Determining suitable PCB track width

Each PCB design is unique and requires the skills of the designer to adapt the design to fall within PCB manufacturing capabilities. One of the key aspects of PCB development is determining appropriate trace sizes for current requirements. This article provides guidelines that designers should follow to determine the trace width.

PCB design is the art of designing a PCB layout by graphically locating components within a CAD package onto a layout and then connecting appropriate pins from one component to the other with tracks. This appears to be a technical process that utilises simple formulas and rules. But, as we all know, in the real world nothing is ever black and white. The PCB layout is usually a compromise between conflicting requirements. These compromises are often multidimensional in nature.

As an example, how do you determine the width of the PCB trace, in order to carry the current imposed by the circuit? This is one of the most basic questions, yet there is no clear formula or answer. The best answer is “as wide as possible”. But this is not a quantifiable parameter. It is not a number that you can enter into the PCB CAD package. Board designers need  numbers that they can work from.

The current carrying capacity of a PCB track is fundamentally determined by two parameters:

• Heat generated by the track. This is a product of the current flowing through the track and the voltage drop across the track (P=IV), essentially the electrical resistance of the track.
• The ability of the track to dissipate the heat generated in itself as a result of its electrical resistance.

The first parameter (track resistance) can be easily calculated and compensated for by increasing copper thickness and track thickness. Copper tracks are resistive. As a result, any current travelling through the track will produce a voltage drop which will cause the track to dissipate heat. Unless the track’s resistive characteristics are specifically required and form part of the electronics design of the circuit, power dissipation by the track is undesirable. As a result, the lower the dissipation and the lower the voltage drop the better. Hence, making the trace as wide as possible should be the design directive.

Temperature rise is used to derive the minimum trace width. One would think from this point that it is simply a matter of determining an acceptable temperature rise for the track in question, entering the numbers into the trace width calculator and presto, there’s the answer. But it is not so simple. The second parameter is more complex as a number of elements come into play. These include air moisture, mechanical orientation (which affects convection airflows), cumulative heating affects from adjacent tracks and components, and copper distribution adjacent to the track and on other parallel layers to the track.

For this reason, calculating the exact current-carrying capacity for a copper track, given a limiting temperature rise for the track, is complex. Most trace width calculators are based on the IPC 2221 (ANSI / IPC-D-275) standards. These standards were extracted from charts produced by the US Department of Defense over half a century ago.

Track spacing should be 20% more than track width, ie, 8 mil tracks should have a 10 mil spacing; 5 mil tracks should have 6 mil spacing; 15 mil tracks should have 18 mil spacing.

To determine the track’s temperature the designer needs to recognise that the temperature rise of a track is dependent on the heat generated by the track - as measured against the characteristic of the track’s ability to dissipate that heat. The ultimate stabilised temperature rise will be the temperature at which equilibrium is established between these two parameters. The heat generated by the track is a simple implementation of the P=I2 x R formula. However, even this is not straightforward. The resistive value of the track is dependent on temperature. The resistance increases with increasing temperature. The trace width calculator resolves this issue. However, the copper thickness can vary substantially from one board to the other. A board specified with 1 oz copper finish can easily have a tolerance of ±25%. In addition, whether that 1 oz of copper thickness comprises 1 oz pure foil or ½ oz foil plated up to 1 oz (which is the norm for double-sided/multilayer build) also affects copper’s conductive characteristics.

The dissipation of the heat away from the track on the other hand is less straightforward. The heat conducted away from the track is determined by the ‘things’ surrounding the track and the thermal resistance to these ‘things’. These include: humidity, and the adjoining copper tracks and fills (both next to the target trace and also on parallel layers). The trace width calculator does not take these considerations into account, nor does it consider the thermal conduction of the solder mask over the track. It also doesn’t take into account cumulative heating effects from adjacent tracks and components. In short, the trace width calculator provides some sort of approximation of power dissipation and provides the engineer a figure that is more of a guess as to the temperature rise above ambient, which the track will experience on a typical PCB.

What does a temperature rise above ambient actually tell us? If the optimum track size (as wide as possible) has been determined, does it matter what the temperature rise above ambient will be. What really matters now is the reliability and the mean time between failures (MTBF). The designer needs to consider how the temperature rise of the track will affect the reliability of the track as well as the components on the board that will be subject to the heat conducted away from the track. This is the primary consideration in determining the track width.

Extremely high temperatures will cause the solder mask protecting the track to break down and fail. Prolonged exposure at high temperature will then cause the exposed track to oxidise and eventually corrode; however, this is the unlikely extreme end of the scale. The temperature rise is more likely to be moderate. The failure mode then presents itself as metal fatigue due to expansion and contraction of the track and of the thermal cycles it has encountered. So we are now moving away from the conclusion that temperature rise is the key element. The other important factor is the thermal stress induced by the magnitude of the temperature rise alongside the timing and number of cycles of the temperature changes. This is extremely difficult to calculate as the geometry of the tracks and the end connection points of the track, along with the coefficient of thermal expansions of the dielectric as compared to the track itself, are key factors. All we are trying to determine is what width we need the track to be and given that the design may contain 100s of tracks, we cannot procrastinate on this one design consideration for too long.

What we are really trying to achieve is to design a product which is as reliable as practicably possible and this is achieved by making the tracks as wide as possible.

Physical constraints and creepage and clearance distances impose restrictions on the maximum track width that can be utilised, and tools such as the ‘trace width calculator’ provide approximate guidelines as to what temperature rises will be encountered. The designer needs to understand that these are approximation tools only and that the real-world elements which determine what the minimum track width needs to be are complex.

In reality, PCB technology has changed very little for over half a century. Tolerances, dimensions and techniques have been improved but fundamentally the technology is the same. Complex considerations discussed above have not imposed significant limitations on the reliability of PCBs because the technology and material used in PCBs is inherently robust and resilient to poor design practices. For this reason a designer needs to use intuition as much as pure theory to determine track widths in order to ultimately produce a high-quality design.