Prepreg thickness and pattern in PCB layers

I have found a table in which prepreg styles vs thickness and the percentage of resin is given. For example in 2116 there is less percentage of resin compared to 106 having higher percentage of resin. What this has to do with percentage of copper ? It says 100 % copper.

Do we have these styles / patterns (106, 2116 etc) only for prepreg or they are also for defining core materials ?

The resin percentage doesn’t have anything to do with copper. Generally speaking, the more resin you have the better because your dielectric constant will be more consistent, smoother if you will.

For 106, Isola offers at least 3 variations: standard, expanded and spread. See the image.

Expanded and spread offer better performance.

106, 2116, etc. also define core materials. We carry both.

1 Like

Thanks for reply. This means that thinner dielectric materials for example 106, 1086, 1080 are better then thicker material 2116. This is because the dielectric constant value is more consistent in thinner material having higher resin content.

1 Like

No I think it is not like this.

Referring to the document “General High Speed Signal Routing” a TI document SPRAAR7J – NOVEMBER 2018 – REVISED FEBRUARY 2023

Here is the link

A text on page 5.

“Because the ratio of fiberglass to epoxy is the primary contributor to the Ɛr disparity, choose a PCB style with a
tighter weave, less epoxy, and greater Ɛr uniformity across longer trace lengths. Before sending your design out
for fabrication, specify a PCB style that can best accommodate high-speed signals. For examples of common
PCB styles, see Figure 2-4.”

Glass style 7628 is shown better then glass style 1080.

Glass style 7628 is thicker and glass style 1080 is thinner.

That means thicker dielectric dielectric materials have tighter weave, less epoxy, and greater Ɛr uniformity across longer trace lengths, correct ?

  1. Type and style of weave have no relationship to thickness. The vendor selects a certain pattern of cloth, soaks it in the epoxy mixture, adds whatever additional epoxy that might be needed, then heats it to set it in place permanently. The type and style of cloth would only have a secondary effect on thickness.

  2. If there were no cloth, the dielectric constant would be completely uniform. It is the cloth’s addition that causes the dielectric constant to vary, and the different styles of cloth determine just how much it varies. Looser, more spread apart weaves allow a little more uniformity. Going with my cake example from below, whole walnut halves provide a lumpy (non-uniform) texture as opposed to grinding up the nuts to a fine powder to achieve a nice, smooth, uniform texture. If you grind them to a fine powder, you get Rogers materials.

  3. The dielectric constant is normally not changed by overall thickness. Say you made a cake and the batter had 10% raisins. Whether you make a single layer cake or a 2 or 3 layer cake, the percentage of raisins does not change (so the cake has a uniform dielectric constant).

I assume that percentage of fiberglass or epoxy representing glass styles in dielectric materials.

Kindly have a look at the attached three pages taken from the TI document.

Glass style 7628 is shown better then glass style 1080.
I guess each glass style refer to one particular thickness, is that true ? I have seen a table in which it is shown that dielectric with a particular glass pattern has a particular thickness.

If yes then please let me know which dielectric glass style pattern is thicker and which one is thinner when comparing the glass style 7628 and glass style 1080.

Or there is no relation between glass style (percentage of fiberglass or epoxy) and the thickness of dielectric material.

Comparing the glass style 7628 and glass style 1080, which one has more uniformity in dielectric constant value ?

Hello Sajal, First, I want to let those on this thread know that the glass suppliers have converted all glass styles to a spread fabric. WE actually have to order the glass non spread if we have a request. Not very common.
Second, the Resin Content % is based on treated weight as processed at our facility. Since the glass and resin have significantly different densities, resin content by volume would be a better way to understand the ratio of resin to glass.
If you are designing a high speed, high layer count PCB, 7628 is most likely going to be too thick and you will need to use thinner glass styles with high resin content.
Consistency of the Dk is determined by thickness control of the dielectric in our treater. Dk is a function of the ratio of resin to glass based on volume. As long as we control the thickness in our process, the Dk will vary very little for long transmission lines. For shorter lines, the glass weave will create micro Dk effects. I am happy to share our glass presentation that shows the available weaves for most products in the industry. The images that are contained show the measurements of the glass weave so you can see the periodicity of the weave.


Thanks for taking this one, Michael!

Deep subject actually. There are a lot of engineers out there that do not have the information.

How do I post a pdf? Or can I?

You can’t. You can share images and URLs only. Do you have a link you can share here?

No link unfortunately

Here it is: Glass Fabric 04_2022.pdf?version=0