Engineering Spotlight: Dielectric Constant Mismatches

By Mark Thompson | Published on May 13, 2009 - 10:00 AM PST

How many times have you had a part go from Prototype to production only to get a whole lot of "door stops"? Inconsistencies between prototype and production type board shops are only the beginning. There are variances between prototype shops themselves, and even shops under the same Company name have Inconsistencies between their own shops that can occur due to anything from press parameters to environmental issues. And whereas I have always been a fan of "the more information the better”, Sometimes too much information can cripple an expedite through a Fab shop.

An example of this is noting critical parameters of the job to a specific value when there is so much variation between Fabricators (both proto and production) that a compromise must be proposed.

Let me give you an example of this, Lets says a Customer researches material properties via a web site for the material manufacturer, Let’s use fr-406 for the purpose of this example. Typically, electrical properties pages describe a specific value for Dk based upon test method and the frequency at which the device will run. In the case of fr-406, anywhere between 4.2 and 4.9 (Higher frequencies have the lower Dk’s).

Now the Customers engineer models/simulates Impedances based upon these numbers. (Let’s says for this example they choose 4.5 Dk) The job then goes through its normal release process, mechanical people are involved and the job is released for prototype fabrication. Herein lays the first hitch. fabricators tend to try to minimize costs, In order to do so a fabricator may choose to build as a foil cap type construction to minimize cost (less material is used, and more pre -preg or B-stage is used, which is cheaper than core).

What does this mean to you the customer? Who has modeled Impedances using Material literature based upon cured core values?

It means, for example, in a Micro strip type application that the interface material between the Intended Impedance controlled trace and its ref plane is now pre preg. The smaller the trace geometries for the microstrip the thinner the distance needs to be to the ref plane, In a Fabrication environment attempting to save costs this means very resin rich pre pregs. In some cases like in the case of 106 the Dk can be as low as 3.3, and If the Customer has modeled at 4.5, this mis-match in Dk's results in having to increase the trace size. Sometimes to the point of where there is not enough air gap (space) left after re-sizing for impedance and this is where the compromise comes in.

Many times, if Air-gap/space is available, the lines can be increased or decreased to accommodate the Impedances. Having said that, a fabricator should really try to err towards altering dielectric distance whenever possible before altering any Impedance /Signal Integrity lines at the Fabrication level, as not much is known about specific performance characteristics of a given signal and we certainly would NOT want to decrease current carrying capacity to accommodate Impedances unless the End User has approved.

The bottom line for this Column today is to consult your Fabricator at the earliest possible time in the process to get specific Effective Dk's of each critical sub section prior to laying out traces. If Multiple vendors may be used for board fabrication, get a sampling of models from a few of them using the same material and compare the pre preg-dielectric values and Dk’s used in each case averaging between them. In this way the Customer has built in 10% deviation allowance and the part can transcend to production environments easier and with less compromises....

Thank you for your time.
Please feel free to call me with any questions or comments at:

(425) 823-7000 extension 239
or E-Mail me at MarkT@Prototron.com