May You Survive HDI PCB

HDI PCB manufacturing pushes boundaries with its fine lines and spaces. Add blind, buried and micro vias to the design and things can quickly go wrong. Here are some tips you will need to succeed in HDI.

Hi PCB designers, today we are going to talk about HDI.

In the past, high-density interconnect has run up against the wall of PCB technology. And today, with lots of innovation and investments, we are able to push the boundaries of PCB fabrication when it comes to trace widths and spaces. So HDI is essentially defined by the spaces between traces as well as the spacing between all copper features. In that regard, we want to define HDI and microelectronics and provide a few suggestions to you on understanding design rules and how they interplay off each other in order for you to get a good design.

Why is there even a separate category of PCB technology called HDI? Well, when the lines get smaller than 65 microns, the ability to etch your trace and space is diminished. You can’t really use a traditional etching process. For example, typical etching on a standard board allows for a very thick resist along with standard imaging on an LDI machine and you essentially have plenty of space for any tolerances needed to etch those spaces. In a microelectronic environment, you don’t have any of those leeways. The features are so close together that traditional etching processes don’t work.

The number one question designers usually ask is… what pad sizes, what hole sizes and how tight can the traces and spaces be together? These are good questions as long as you’re designing a relatively standard board. You can’t just look at our HDI PCB manufacturing capabilities page and not understand how they interact with each other. For standard products, it may work but as you get into more complicated boards such as HDI or microelectronics with finer feature sets it may not work.

Let me give you an example…

Let’s say you’re doing an HDI construction with very fine lines and you are going to do, let’s say a 2+4+2 type of stack-up. So 2+4+2 essentially means you’re going to have a buried sub-construction which we will mechanically drill and plate. Now if we have to do fine lines on layer 3 of the buried sub-construction, that becomes problematic. The processes on layer 3 outer layer don’t allow for fine line technology. Whereas on layer 2, you have more leeway
for your fine lines.

Now when you get into that feature set, designing the whole board with those feature sizes is not the same as designing just a small portion of the board with those feature sets. If you can loosen up the spacing with bigger pad
sizes and bigger drill sizes in non-dense areas of your PCB design it will make it easier for the manufacturer and less costly for you… Especially if you’re going into production.

So I really wanted to dive deep on that distinction. Even though our capabilities would say 35 microns is the minimum trace and space, it is not a feature size that we would recommend or be able to manufacture across the entire design.

There is no such thing as universal design rules.

Let’s say you’re connecting a micro BGA to a connector and you have to route those connections. The board is 3” by 3” so there’s a lot of nets you need to connect in a small space. Is it wrong to design a 4-layer board with 1-mil trace and space or 35 microns just connecting all the nets? Yes, it is. So how do you attack this from a design and manufacturing perspective? There are 2 stack-up decisions you need to make:

1. How many routing layers are you going to add?

2. How are the routing layers going to be connected?

You should always try to create a stack-up with no less than 3-mil trace and space routing density. If you need smaller trace and space, it is not a problem. But start with the 3-mil in mind. Now if you do that… something has to give. Either your board has to be larger or there needs to be more layers. So if there’s space constraint and you need to maintain a form factor then you’re going to be adding more routing layers, which is absolutely fine. Now how are you going to connect these routing layers? Are you going to connect with through-hole vias? Which basically “swiss cheese” all your layers.

Or are you going to be able to connect with blind and buried vias that are created by laser drill and free up a lot of that routing space for you?

Option 2 is a much better choice than option 1.

And in many cases, option 1 is not even possible. Laser-drilled vias require a much smaller pad size which helps you design faster. And there are plenty of PCB fabricators that can manufacture HDI boards. So once you figure out your minimum routing density required, then you figure out your stack-up, and where your vias are going to start and stop. Sierra Circuits provides a stack-up tool that really gives you a good starting place as to what is a manufacturable design for HDI and what your technology levels for trace and spaces along with pads and vias.

Other considerations for HDI are what surface finish you choose, what copper thickness you pick, the impedance values and material types. Try our free HDI tools such as the Stackup Planner and Material Selector. For more information, download our HDI design guide.
HDI Design Guide

Tags:

Leave a Reply

Your email address will not be published.