Engineer experience: How to measure high precision DC voltage?

A colleague once asked, “How do I measure microvolts during testing?” High-precision DC voltage measurements can be complex. Time is money during the measurement process. Therefore, achieving fast and accurate measurements has always been a challenge.

Traditional optimization techniques use high precision amplifier circuits and faster measurement devices. To achieve the best measurement in the shortest possible time, the above two are still necessary, but not enough. The inverse relationship between the stable delay time and the signal noise depends on the equivalent noise bandwidth of the measuring device drive circuit. The device under test (DUT) and the measuring instrument define system characteristics that closely tie the stable delay time to the broadband noise.

If the circuit bandwidth is zero and the noise will be zero, we may use a sample to make the measurement, but the circuit will never be stable and the DC error will be 100%. Therefore, too low a bandwidth will cause the measurement time to be too long. On the other hand, if the bandwidth of the circuit is infinite, the stable delay time will be zero, but the broadband noise will also be infinite, so we lack sufficient measurements to average. Thus, the faster the amplifier speed, the longer the time required for voltage measurement with high precision may actually be reversed.

Let's explore this relationship. In the test sequence, the output of the DUT must settle within a predefined error range after a voltage step. Assuming a unipolar step response, the settling time will depend directly on the bandwidth.

Each voltage measurement contains broadband noise generated by the DUT, amplifier, and resistor. The amplifier produces voltage and current noise; the resistor produces Johnson noise. Since the roll-off characteristics of the filter are not infinitely steep, the noise becomes less important in the region after the –3dB cutoff. The effective noise bandwidth takes into account the noise in this area.

If the broadband noise and effective noise bandwidth are fixed, the number of samples required is determined by the measurement tolerance. The basic statistics give the average number of samples needed to achieve a 98% confidence level when the amount of noise is constant. This deviation of the average reflects the repeatability of a single DC voltage measurement. There are many problems with high-resolution measurements, and this article cannot be exhaustive. Below we will discuss the importance of solving the overall problem.

Stable delay time. If there is a settling time problem with a component in the circuit, the overall measurement time is increased. Limited slew rate is a common cause. It is always calculated using small signal stabilization time. Dielectric absorption is a detrimental problem, so care must be taken to select the filter capacitor.

Stabilize the goal. The setting of these target values ​​is easily too small, such as 0.0001%, and the result is a dynamic increase in measurement time. Since the target is affected by the step size, a larger target should be used when the step size is a small fraction of the measured dynamic range. It may be necessary to set the bandwidth separately for different measurement sequences.

Error voltage. For all measured values, the allowable error voltage is often set too small. Statistics show that if the Student T table value of 1.6 is used, the measurement deviation will be within the tolerance within 98% of the time.

Voltage reference. This can introduce noise. In the case of a D/A converter, these noises may be related to the code.

Broadband noise. Use a high-quality spectrum analyzer to directly measure the broadband noise of the circuit. With the same number of typical circuit noise sources, accurate calculations with paper are quite tedious and error prone.

Measurement accuracy and resolution. Test engineering practice generally requires that the resolution of the measuring device be greater than the tolerance, but in fact always assume that the accuracy and resolution of the measuring device is much smaller than the tolerance in the actual measurement.

Amplifier. Use a low noise op amp in the signal chain. This is a good way to keep the resistance low, but not so low as the amplifier generates current drive and thermal problems.

Test cost requirements require optimization of traditional slow, high-precision measurements. This technology allows us to minimize measurement time, save money, and is an attempt in test design.

The semiconductor industry is at the turning point in the production of 20-bit DC circuits. The next question is the need for a test engineer with good professional skills.

 Light reflectors in China with resonable price and good quality,We hope to establish cooperative relationship with you.

Light Reflector

Light Reflector,Aluminum Light Reflector,Street Light Reflector,Energy Save Light Reflector

Yangzhou Huadong Can Illuminations Mould Manufactory Co., Ltd. ,