All 50Ω symmetric striplines in FR4 have the same aspect ratio. Using a 2D field solver, we can explore what this ratio is for a specific value of Dk and conductor thickness. Spoiler summary: The ratio of the line width to the dielectric thickness of a 50Ω symmetric stripline in FR4 is about 0.8. This is a simple to remember, easy to use consistency check for your designs. In the last rule of thumb, we showed that all 50Ω microstrips in FR4 had the same aspect ratio. If you double the line width, you have to double the dielectric thickness to maintain a 50Ω line. In this respect, all 50Ω microstrips in FR4 look the same. This has two important consequences. The first is the obvious one. A simple rule of thumb is that a 50Ω microstrip in FR4 has an aspect ratio of line width to dielectric thickness of 2:1. This offers a quick and easy way of estimating the stack up conditions for surface traces. The second consequence is that crosstalk will also scale with line width. This makes it so much easier – at a glance – to evaluate the crosstalk influence between surface traces. Crosstalk is driven by fringe fields. Fringe fields are driven by the dielectric thickness. Increase the dielectric thickness, and the fringe fields extend farther, and crosstalk, for the same spacing, increases. But when you look at surface traces, you do not see dielectric thickness; you see line width. Knowing that line width scales with dielectric thickness means that crosstalk will scale with line width. This is the basis of Rule of thumb: Line spacing for near end crosstalk: What spacing, in terms of line widths, is required for a given worst-case crosstalk in microstrip or stripline. We assumed 50Ω lines in FR4. In the same way as with microstrip, we can explore the aspect ratio – the line width to dielectric thickness – for a 50Ω symmetrical stripline in FR4 using a 2D field solver. In this example, I am using the Polar Instruments SI9000 tool. We select a nominal Dk of 4.2 and half-ounce trace thickness. We assume the same Dk for the core and prepreg layers, and for a symmetrical stack up, the same dielectric thickness, H1 and H2, as shown in figure 1.

EDNAOL 2016MAY20 EDA TEST TA 01Fig1 Figure 1: Cross section of the symmetric stripline structure.

As we change the dielectric thickness of both the top and bottom layers, we calculate what line width is needed for a 50Ω line. Then we plot the aspect ratio for a 50Ω line over the range of dielectric thicknesses being evaluated. The result is plotted in Excel and shown in figure 2.

EDNAOL 2016MAY20 EDA TEST TA 01Fig2 Figure 2: Aspect ratio for a 50Ω, symmetric stripline calculated with Polar Instruments SI9000 2D field solver.

Looking at the graph, we can state RoT: Aspect ratio for 50Ω stripline:

For line widths around 5 mils, the aspect ratio for a 50Ω symmetric stripline is about 0.8. Even as line width increases to 15 mils, the aspect ratio increases less than 10%. Not bad for a simple rule of thumb.

This means that for a 50Ω symmetric stripline in FR4, the dielectric thickness on the top and bottom of a 5 mil wide line is about h = 5 mils/0.8 = 6 mils. Knowing the aspect ratio for these specific conditions, we can also estimate the directions for adjusting this simple rule of thumb for other conditions. If the Dk were lower than 4.2, the dielectric thickness would have to decrease to maintain 50Ω. If the conductor thickness increased to one-ounce copper, the dielectric thickness would have to increase. Just how much to adjust the dielectric thickness as other parameters change is hard to estimate with a simple rule of thumb. If it's important to know exactly what the stack-up dimensions are, don't use a rule of thumb; use a 2D field solver.
About the author
Eric Bogatin is Signal Integrity Evangelist at Teledyne LeCroy.