You are here:

Views: *285*
Author: *Site Editor*
Publish Time: *2023-03-22*
Origin: *Site*

The length, width, and thickness are chosen naturally when deciding the board's dimensions. The device's size determines the height and width, so specify the dimensions unless the device has a lot of room. The thickness should be 1.6 mm, the industry standard, or as thin as possible for a light, thin, short, and small device.

Sometimes you glimpse the finished product and don't think about the cross-sectional dimension of the board, that is, the cross-sectional view, without considering it from the beginning.

If you carefully consider the cross-sectional dimensions of the port, you can achieve what he used to use, a** 6-layer board** with his 4-layer board. Let's start by taking a closer look at this cross-section.

Some parameters are not freely selectable, but let's consider each one.

A typical** board thickness** of 1.6 mm is the industry standard. If it deviates from this standard, depending on the lead length of the inserted part, the lead may not protrude on the solder surface, or conversely, it may protrude too much, requiring a lead-cutting process.

Determining the standard thickness is a prerequisite for memory modules, where the board becomes a card edge connector.

Naturally, the board for portable devices is much thinner than the standard 1.6 mm. In addition, it is necessary to consider the strength and warpage of the board at the same time.

A prepreg material is used to adhere and keep distance when laminating each layer. It is a sheet of gauze-like fibre made of glass impregnated with semi-cured epoxy resin.

Since it has thermosetting properties, it can be cured by stacking layers and applying heat and pressure. The thickness is kept constant by the thickness of the glass fibre. Still, since the thickness of this glass fibre cannot be freely and continuously selected, for example, if two sheets of 50 um are used, the thickness can only be chosen discontinuously, such as 100 um. Yeah.

Generally, the conductors' thickness in boards is 18 um or 35 um. It means that 1 square foot weighs 1 ounce. The former is the 1/2-ounce copper foil, and the latter is the 1-ounce copper foil.

Although 18 um copper foil is generally used for the surface layer, through-hole plating is applied, so the finished conductor thickness is about 40 um.

Although it cannot be said unconditionally because it changes depending on the circuit constant, we consider the crosstalk coefficient ξ (ξ: Kusai, a lowercase Greek letter) as a measure of crosstalk. ξ is typically on the order of 0.1 to 0.3.

As I mentioned in the specialist column " How to determine the characteristic impedance of the board ", it seems that the characteristic impedance of the port is usually selected to be 50 to 60 Ω. "Apparently" doesn't have to be this value, and there may be options such as 30 to 40 Ω or 70 to 80 Ω.

Choosing a low value eliminates the need for a damping resistor and reduces crosstalk noise. However, since the amplitude (voltage) of a wave propagating on a line is the product of the characteristic impedance and the propagating current when the characteristic impedance is lowered, the current increases and the simultaneous switching noise increases.

Conversely, if you choose a high value, you can reduce the simultaneous switching noise, but you will need a sizeable damping resistor, and the crosstalk noise will increase. I think 50 to 60 Ω was chosen as a compromise between the two because it is easy to make.

Characteristic impedance and crosstalk coefficient are key parameters when determining the cross-sectional dimensions of the port.

It is important to determine these parameters from the beginning rather than start the design by obtaining them from the finished dimensions. Determine this value first, then vary the pattern spacing and distance from the ground to determine the optimum dimension.

Consider a surface layer with characteristic impedance Zo of 50 Ω and a crosstalk coefficient ξ of 0.15. The dielectric constant εr (ε: epsilon in lowercase Greek letter) varies depending on the board manufacturer, but in many cases, εr = 4.7 is chosen, so this value is used.

· The pattern width is W.

· The distance between the patterns (gap) G.

· The distance from the ground h.

· The pattern thickness t.

In the case of the surface layer, the thickness of the solder resist that covers the pattern also affects the characteristics, albeit slightly. Here, we assume a solder mask that covers 10 um above the pattern. Of these parameters, h is determined by the prepreg thickness mentioned above, so it cannot be freely selected continuously.

First, as the first step, determine the pattern width W of a single line and the distance h from the ground. The cross-section of the pattern has a slightly smaller trapezoidal shape due to the etching effect.

Zo is 50 Ω for the rectangular case and 50.6 Ω for 116 um upper and 126 um lower bases, so the difference is only about 1%.

Even if the accuracy is sought and calculated at this examination stage, the board manufacturer only uses the result as a reference. Corrections are made according to the manufacturing process, so please consider the fine accuracy meaningless.

On a normal board, W ≥ 100 um, so if h is a multiple of 50 um, then h = 100 um and W = 126 um. Currently, if ξ = 0.15, then G = 155 um. We want to select W as small as possible, so if W = 100 um as in (2), Zo = 55 Ω, and when ξ = 0.15, G = 162 um. Zo is 10% higher than initially set, so we have to accept that, but there is a solution to increase the damping resistance slightly.

It is a solution that does not increase Zo. Another solution is W = 100 um, G = 131 um, h = 80 um if the distance h from the ground can be a multiple of 40 um instead of a multiple of 50 um.

It is an example of setting parameters without overthinking, such as 200 um for h when using a 1 mm core material with a thickness of 1.6 mm.

The area column shows the area occupied by one line when normalized. (2) and (3) are slightly improved, but if the parameters are determined without thinking, the area will be 3.7 times larger.

Table 1 v shows when t is chosen to be 9 um. If the pattern thickness becomes thinner, it will be possible to make the pattern finer and the gap more negligible, so W/G = 75/75 um may be achievable. The occupied area at this time is 28% of I, and the crosstalk coefficient ξ is smaller than i.

Copper foil with t = 9 um has been used as UTC (Ultra Thin Copper). The pattern on the surface layer becomes thicker due to plating, but this can be achieved by selective plating that applies plating only to the through-hole portion or by scraping off the plating applied to the entire surface by etching or polishing.

If the mounting density of only the patterns is tripled, the 8-layer board becomes a 6-layer board, and the 6-layer board becomes a 4-layer board. I can't. It's worth considering once.

If you compare (a) to (c), the difference in mounting density will become clear, so please consider it carefully in advance.

The ounce is usually 28 grams, but gemstones and precious metals are called troy ounces and weigh 31 grams. In the case of copper, which one is used? If the specific gravity of copper is 8.95 and 1 foot is 30.48 cm, it is 29 g at t = 35 um. g) cannot be definitively determined. Originally, the SI unit system should be used, and the yard-pound system should be expressed in parentheses, such as 35 um copper foil (1 ounce) instead of 1-ounce copper foil.