1 Experimental Procedure

The procedure used to determine the viability and performance of using a copper trace as a shunt resistor involved testing several different copper trace widths, lengths, and geometries. The experimental setups are shown in Figure 1-1 and Figure 1-2. These two boards are collectively referred to as Revision A.

In total, four trace widths were tested: 8 mil, 100 mil, 200 mil, and 1750 mil. Each of these traces was tested at one, two, and three inch lengths with current flowing through them. In addition, each trace was tapped from the bottom of the trace (lower edge of the trace on the red layer) and from the center through vias to another layer (blue layer). Tap off points in this document refer to locations on the copper traces across which the voltage is measured using a differential amplifier. If the temperature and dimensions are known, the resistance between the two points can be calculated. Tap off points were chosen in the bottom and in the middle of the traces to investigate the impact of sense line location on the voltage measurement. The amount of current flowing through a conductor varies inversely with the resistance of the wire, meaning that measuring voltage in the center should theoretically yield a different result, since current would have to travel through more copper and therefore more resistance to reach the bottom of the trace as opposed to the center. Testing multiple lengths by using different tap points on the trace demonstrates the linearity of the trace. If the trace resistance behaves linearly, the three-inch tap off points should produce a resistance three times that of the one-inch tap off points. Also visible in Figure 1-1 is what is referred to as the square shape geometry (the trace second from the top). This is an 8-mil trace that is intended to examine the effects of changing the path of current on the overall resistance measurement.

The copper trace shunt tap off points are connected to an INA190 instrumentation amplifier through copper trace sense lines. The INA190 series of devices are voltage-output current-shunt monitors (also called current-sense amplifiers) that are commonly used for overcurrent protection, precision current measurement for system optimization, or in closed-loop feedback circuits. This series of devices can sense drops across shunts at common-mode voltages from –0.1 V to +40 V, independent of the supply voltage. Five fixed gains are available: 25 V/V, 50 V/V, 100 V/V, 200 V/V, or 500 V/V. The INA190 was chosen for its low input bias current, which increases the accuracy of the voltage measurement. For each trace, the INA190 output was recorded at each point (one, two, and three inches stemming from the bottom and from the center of the trace) at several known currents and using the current and the gain value of the amplifier, converted to a resistance measurement. The maximum current level to be pulled through each trace was determined using a current width calculator from the manufacturer. Four separate currents (1 A, 5 A, 10 A, and 50 A) were used to calculate the minimum trace widths to stay within a 20°C trace temperature rise. These widths were rounded up to obtain a trace width that could safely handle the given current, and these currents were in turn defined to be the maximum current that a particular trace would be tested with. The 8-mil trace takes 1 A, the 100-mil trace takes 5 A, the 200-mil trace takes 10 A, and the 1750-mil trace takes 50 A. INA190 gain settings were chosen to avoid saturation of an ideal copper trace resistor with a given maximum current flowing through it, assuming a supply voltage of approximately 5 V. The 100-mil, 200-mil, and 1750-mil traces all used the INA190 A3 with a gain of 100 V/V, while the 8-mil trace used the INA190 A2 with a gain of 50 V/V.

Calculating the resistance from a given INA190 output made it possible to see how the copper trace resistance changed with increasing current levels. Additionally, these tests were repeated in a temperature-controlled environment at four different ambient temperature levels: 0°C, 25°C, 55°C, and 85°C. The 25°C setting in this case, while conventionally taken to be room temperature, differs from what is referred to here as room temperature in that 25°C was the temperature controlled value, and room temperature refers to measurements taken out of the temperature-controlled environment. Testing at four different temperature levels gives an idea of how the resistance changes with temperature. For these tests only the three-inch test points were used, but once again both the center and bottom tap offs were tested.

Current step measurements were obtained by adjusting an electronic load in series with the trace to pull the desired current level. To achieve precise results, the voltage across a precision resistor in series with both the traces and the electronic load was measured using a multimeter to accurately determine how much current was moving through the trace. Each trace was tested initially at 0 A as specified by the electronic load, although the actual current value flowing through the trace as measured by the resistor and multimeter combination varied across a small range of positive and negative current values. The common mode voltage was set to approximately 5 V.

Although not visible in Figure 1-1 and Figure 1-2, a piece of thermally-conductive tape was used to attach the TMP235 temperature sensor to the middle of each trace. The TMP23x devices are a family of precision CMOS integrated-circuit linear analog temperature sensors with an output voltage proportional to temperature. This sensor was covered in thermally-resistive tape to further increase contact with the trace and insulate from the surrounding environment. This sensor obtains data about the temperature of the trace to determine how temperature affects trace resistance and performance. Temperature data obtained at the 0 A current step was used as a baseline temperature measurement and is intended to account for any variables that might cause discrepancies between the trace temperature and the ambient temperature, especially in the more controlled environment.

Results are grouped by whether or not they were obtained in a temperature-controlled environment. The process to obtain data is relatively straightforward. For room temperature measurements, the procedure involved incrementally stepping from 0 A to the maximum current level for each trace following a roughly logarithmic scale. Upon moving to the next current step, INA190 outputs were recorded for the one-, two-, and three-inch locations. Then, using the TMP235EVM evaluation software, temperature output from the TMP235 sensor was monitored until it ceased increasing. At this point, 50 temperature samples were recorded at a 2-Hz sampling rate and averaged to determine the trace temperature. During the course of testing it was observed that a potentially non-trivial temperature rise occurred between the last tap off tested and the recording of temperature, meaning that INA190 outputs could have possibly changed during the temperature stabilization process. For this reason, the procedure was modified slightly for the temperature-controlled tests. Since these tests took place in a temperature chamber, the TMP235 was first used to make sure that the trace had equalized with the ambient chamber temperature before any testing began. Then, a reduced number of current steps were used to again sweep from 0 A to the maximum. However, in this case, no values were recorded until both the INA190 output and the TMP235 output had stabilized. Results from the two testing setups are detailed in Section 2.