How to measure pulse waveforms using high-speed A/D converters

As the array of products that use pulsed signals continues to grow, including today's more energy-efficient ICs, switching power supplies and inverters, and even LED modules and sub-assemblies, the measurement of the discrete components of these end products under pulsed conditions has become extremely important. Test instruments with only DC source output capabilities can apply enough power to the device to generate enough heat to change the device's characteristics. The use of pulsed excitation signals also requires instruments that can achieve faster measurements.

Comparison of High Speed ​​and Integrating ADCs

Traditionally, precision SMUs (source measure units) have used integrating analog-to-digital converters (ADCs), which average signals over a certain time interval (called the integration time). Figure 1 depicts a simplified dual-slope integrating ADC, which works by charging a capacitor with an unknown signal and then discharging it at a reference voltage. The ratio of the charging and discharging times is proportional to the ratio of the unknown signal to the reference signal. While this ADC technique provides high accuracy and excellent noise immunity, the capacitor charge-discharge cycle results in long measurement intervals (at least 50µs), which greatly slows down the measurement. In contrast, high-speed ADCs can sample signals at burst rates up to 1MHz. Unlike integrating ADCs, these high-speed ADCs use sampling techniques similar to oscilloscopes, which take snapshots of signals that change over time. They can provide higher resolution than oscilloscopes (18 bits versus 8 bits, respectively), allowing more accurate transient characteristic measurements with comparable bandwidth.

Figure 2 shows the difference between the results obtained by integrating and high-speed ADCs. Although a high-speed ADC can return more readings, the accuracy and repeatability of these measurements are lower than those made with an integrating ADC. Applications that require higher throughput rates can tolerate lower accuracy or improve accuracy by averaging several readings. In general, measurements made with an integrating ADC with an integration rate of 0.01 PLC or higher can achieve accuracy equivalent to that achieved with a high-speed ADC. Newer SMU designs that integrate two high-speed ADCs can make voltage and current measurements simultaneously. When using these techniques, the combination of high-speed ADCs and advanced triggering modes can support accurate time-varying measurements of pulsed signals. For example, Keithley's Model 2651A High Power System SourceMeter instrument can make measurements asynchronously to the source operation, such as before, during, or after a pulse.

For some applications, such as thermal impedance measurements of power diodes and LEDs, it is important to obtain the slope of the voltage curve at the top of the measured pulse. This function is also useful for measuring the flatness of the pulse amplitude. When the measurement is synchronized with the signal source, a high-speed ADC can digitize the top of the pulse (Figure 3a).

Asynchronous triggering is very effective for spot-average measurements made at the top of the pulse (Figure 3b). Analysis software is often used to average the sampled data to improve accuracy, but newer SMU designs offer average and median filters that can be applied to the high-speed ADC readings to return to spot-average measurements.

Sometimes it is also interesting to measure the transmission characteristics of a pulse as it passes through a device or system. These applications require digitizing the entire pulse, including its rising and falling edges (Figure 3c). This can be accomplished by using a high speed ADC to make the measurement asynchronously to the source operation.

Sometimes a pulse can be used to provide power stress to a device. In these applications, it is useful to record the state of the device before stressing it. This can be achieved by programming a pulse with a non-zero idle level and triggering the measurement before the trigger pulse (Figure 3d). The user can specify how long before the pulse that the measurement should be initiated. A timer can be used to program the start of the measurement and the start and end points of the pulse.

When using pulse testing to stress a device, it is also necessary to measure the device characteristics after the stress is applied. This is generally done by outputting a predefined test voltage or current after the pulse arrives (Figure 3e). The test level should be chosen so as not to cause any additional thermal or electrical stress to the device. The measurement can be implemented by the signal source outputting a pulse with a non-zero idle level while using a high-speed ADC to perform the measurement. The results obtained from the high-speed ADC indicate how the device recovers from the stress.