PCB Cost Drivers

There are quite a few drivers in the bare PCB cost. From the region where the PCB is manufactured, the volume ordered, what features it has, and many other considerations. Below are some estimates for PCBs manufactured in volume in a market outside North America. There will be cost variations between fab houses, and these numbers don’t take into account one-time costs such as NRE, stencils, etc. In fact, these estimates might be outdated. For that reason, these numbers are for developing intuition only and should not be used to calculate pricing! In my experience, the best way to understand how different parameters will affect the cost of your boards is to discuss with your fab shop. 

When I started working on PCBs, I knew a couple of things: Additional layers can help signal integrity and routing, and additional layers cost more. Fantastic! Now just for some hard data. Inquiring to cost, the general advice I received was “more layers cost more,” with no specifics on pricing. Looking at PCB cost calculators from various vendors, why is there such a price discrepancy between them? The dominating factors seem to be market location, material, and turn-around time. 

This write-up will focus on the rules of thumb I’ve developed for estimating cost increase due to adding layers, internal and external copper weight, hole size and density, and material selection. The assumptions that go into these rules of thumb are that this is an estimate only and will vary between fab houses, these estimates are for volume production purchased in markets outside North America. 

I covered how to maximize panel utilization here. While the writeup doesn’t indicate relative costs, it’s still worth mentioning as higher panel utilization leads to lower the wasted cost of the stock material. It’s a good idea to think of panel utilization early, as it can be hard to justify changing the PCB size to increase utilization during a cost-reduction phase if the mechanical elements are defined, and changing the panel size requires changing tooling! Given the simplicity of this step it’s worth considering early. 

More processing, more cost, more labor. Makes sense to me. The tables below show the rule of thumb I apply when adding a layer. The assumption is all things equal when increasing layers; same size/shape, hole size, finish, copper weights, etc.

In the table above, the left shows additional layer costs when adding two layers to the board (e.g., going from a 6-layer board to an 8-layer board adds ~35% to the cost of the PCB). The right shows going from some number of layers (left, red) to some number of layers (top, blue). For instance, going from a 4-layer board to a 10-layer board increases the price by ~194%. 

The graph below shows the same data as the tables above, just in a “cost multiplier” instead of a percentage increase. Starting with a 2-layer board with a unit-cost of 1, this graph plots the cost multiples of adding layers with the above assumptions.

Note: This is PCB cost increase; to calculate the decrease, one must reset the reference point and recalculate. For instance, going from 2 layers to 4 layers is a ~50% increase. However, removing layers, say going from 4 to 2 layers, is a ~33% decrease. 

The higher performance of the material, the more you pay. For this example, I focus on the loss tangent (Df), which is performance in terms of signal loss. There are other factors to consider, such as the materials' thermal properties (Td, Tg, CTE, k), the materials' relative permeability (Dk), and how constant the permeability is. 

Below is a graph showing some common materials and their associated cost increase against their loss tangent. I start with 370HR as I consider that to be a fantastic general-purpose material that is also cost-effective. 

Granted, some materials don’t follow the general trend of lower loss tangent-higher price, but, the trend generally holds. Understanding the performance needs of the system can mean significant cost savings in the material chosen. 

While it’s pretty obvious more copper costs more, I was surprised to find that in some cases, less copper costs more as well! Though this makes sense when taking yield into account and extra processing required due to impacts a light copper pour might have on a thin trace. 

The table below shows my rough approximations of changing copper weight on internal and external layers of the board. The table uses a 0.5oz pour as reference, and the percentages shown are from 0.5oz to the indicated pour weight. 

As one decreases hole size, smaller drill bits are used. Smaller drill bits break more easily, causing the cost of drilling to increase as the operator has to stop the press to change the bit. Increasing hole density past a certain point requires an increase in precision necessary to prevent shorting signals together. The tables below show estimates of cost increase due to decreasing finished hole size, and estimates of cost increase due to increasing hole density in 2-layer and 10-layer boards. 

Not shown, but another factor to consider is the number of hole sizes in your drill table. Changing hole sizes means changing bits, increasing processing time and labor costs. There are some reasons to have different hole sizes (mounting holes don’t typically have a FHS of 10 mils). Still, quite a few times, I’ve gotten a board back from layout to find a long list of hole sizes, and after review, many of the hole sizes could be consolidated with minimal impact on the design. 

Color isn’t a big driver, but Green seems to be the most cost-effective color choice. If you want red or blue, add ~4%. Black or white, add ~7%. 

As with every rule of thumb, always remember Rule 0! These rules provide intuition into a more complex issue to help with early design decisions. Meticulously calculating everything at the beginning leads to a long design cycle. Still, there are some general choices one can make early in the design process that can significantly impact the final cost! 

These numbers were derived based on one individual’s experience in speaking with fab shops and analyzing quotes, are specific to volume production, are specific to PCBs manufactured in a geographical region other than North America, and may be outdated! They’re for intuitive understanding, not cost estimation!