Smooth copper isn't always what it is purported to be.
I always trusted datasheets. Why shouldn't I? Every time I picked one up, I understood they were harbingers of truth about what something did and how to interact with it. From chips to components to widgets, manufacturers have a vested interest in ensuring users know what and how to use their products.
My trust in datasheets was eviscerated the very week I entered the printed circuit board industry. I was attending a weeklong training session on signal integrity taught by Eric Bogatin when he surveyed the attendees about datasheets and their accuracy. While my memory of the day and exact content isn't perfect, I do recall that as the very moment I learned the harsh reality: Datasheets are sometimes not as accurate as one would hope! (As an aside, Eric Bogatin's books on signal integrity are excellent resources because he has that special gift of teaching and explaining complex topics in approachable and understandable ways. The concept of "being the signal" and walking down a transmission line is just fantastic.)
With that established, I will backtrack a bit and say datasheets indeed can be trusted, but it is important to understand the capabilities of the test method being used, the reason why it is used, and the terminology involved.
For the PCB designer and fabricator, laminate datasheets are a good starting point for understanding how material will behave before, during and after fabrication and assembly. While the fabricator is primarily concerned with the mechanical properties, such as glass transition temperature, peel strength, etc., a designer is likely concerned with signal integrity and as such will be focused on a few key parameters:
- Roughness profile of the copper
- Dielectric constant/permittivity (Dk) of the resin/filler/glass combination
- Dissipation factor/loss tangent (Df) of the resin/filler/glass combination
- Glass weave for skew effects.
For the moment, we are going to focus on the copper roughness. As a designer, if you start with a material datasheet, you might see a side column section describing the available copper roughness types. There are no standardized industry terms, unfortunately; only marketing terminology to describe roughness types. There are some common ones, though, but how can one hope to comprehend the roughness profiles when the variety of marketing terminology gets extreme: reverse treat foil (RTF), very smooth profile (VSP), very low profile (VLP), hyper very low profile (HVLP), almost no profile (ANP), etc. These can get confusing fast!
Following Bogatin's approach, let us "walk the path" with a sheet of copper foil, freshly manufactured, to help understand what happens when we read about a 5-mil copper-clad core (CCL) with HVLP (or choose your preferred roughness term). Keep in mind, this journey is meant to convey the concepts, not necessarily the exact processes and order at each step along the way.
We start our journey as a copper sheet just after being created. After all, we wouldn't have been awake yet for being pressed through giant rollers or chemically built up through electrodeposition. As this copper, fresh from the fabrication process, we will be tested for quality control and be assigned to certain lots. The methods for measuring the profile are either mechanically or optically performed, and the results presented as averages for the lots of copper foil types with the methods used on datasheets.
Boom, shuffle, shuffle, screech.
Next, we are arriving at a laminate supplier, which takes our nice "perfect" sides and processes them. Via a process conceptually similar to what the PCB fabricator will do for the entire board, the copper is bonded to the resin, glass and filler that make up the laminate, resulting in the copper-clad laminate (CCL). Now our sides aren't exactly what they were when we left the copper factory, and the inner portions of our sheets of foil attached to the outer sides of the resin, glass and filler are no longer accessible for measuring. But at least they won't change anymore!
Boom, shuffle, shuffle, screech.
Now we are arriving at a printed circuit board manufacturer which is going to take the CCL that we are part of and process us further. We will be subjected to harsh etching chemicals to tear away sections of unwanted copper, leaving behind only the desired traces and planes. A plating process may add more copper for our vias or increase us to the desired thickness. Along the way, we will be processed with some form of oxide, oxide alternative or micro-etch treatments to 1) improve adhesion to the resin, filler and glass combination that is the b-stage laminate (prepreg) and 2) prevent corrosion. Unusual, to be sure, but a horror story I once heard involved a fabricator taking sandpaper to a sheet of copper to "improve" its roughness profile for mechanical adhesion.
Together with other CCLs and prepregs, we are put in a press and heated. The prepregs get gooey and stick to us, cool and harden. At this stage, quality control tests will be performed to verify the mechanical and electrical properties of the board are in compliance with the customer's specification. Once passing, we are done and ready to be sent to assembly where components will be soldered on to us.
So, to review the journey, the copper has been through multiple steps altering the surface at different locations and has resulted in a unique surface profile that we no longer have full access to measure. The only options now are to take cross-section cuts of the board and make side profile measurements of that slice or to make electrical signal integrity measurements of the impedance and scattering parameters.
A lot of thought has gone into correlating mechanical surface profiles to electrical characteristics over the past 70 years, from the Morgan model in 1949 to newer methods like the Huray surface roughness model. But these simulation methods require good data to match with the final board electrical measurements. Since passing from the copper supplier, we have had two different manufacturers make modifications to the surface profile. Do those coppers still have the same roughness profile as our initial datasheet? Hint: Not likely!
If you are a designer who cares about the surface roughness of the copper on your boards, talking with your PCB fabricator to get an idea of what kind of surface roughness you can expect for your final board is important. You may request and pay for some of the smoothest coppers to be put into your board, but if you aren't paying attention to what your fabricator is doing with it, you might just be throwing money away.
GEOFFREY HAZELETT is a contributing editor to PCD&F/CIRCUITS ASSEMBLY. He is a technical sales specialist with more than 10 years' experience in software quality engineering and sales of signal integrity software. He has a bachelor's degree in electrical engineering; This email address is being protected from spambots. You need JavaScript enabled to view it..
