# Design considerations the analog signal chain (Part 2 of 2)

(Part 1 looked at basic analog input such as from thermocouples and RTDs, and some error sources.)

Considerations for measurements in millivolts

Up to this point, we have discussed challenges faced in making precision analog measurements in generic high-precision systems where the range of measurement is typically in the volts range. There are systems that have measurement range in the millivolts (mVs), and these pose a completely new array of challenges for designers.

A good example of such a system is a load cell. Load cells are sensors that convert force applied on them into an electrical quantity. They are generally used in digital weighing scales for weight measurement. A typical load cell is four resistors arranged in a whetstone bridge network. The sensor is excited from the two excitation terminals using a voltage and, based on the force that was applied to the sensor, a small voltage gets set up in the measurement terminals. The output voltage range of a load cell is generally specified in mV/V, which is the output range for a 1V excitation voltage.

Let us take an example of 2mV/V load cell, measuring a maximum weight of 10kg. If the user excites it using a 5V input, then the net output voltage range is just 10mV. This means that even when the load cell is experiencing a 10 kg force, the output would be just 10mV. To resolve close to 16 bits of precision in this 10mV range means that we need to reduce the ADC range to fit this input range.

The most commonly used method is to implement a gain stage to amplify the input to fit the ADC range, thus resolving more bits inside a smaller range. For example, to have a measurement range of 10mV as discussed earlier using an ADC that generally has a 0±1V range, the user can resort to getting close to 100x gain on the signal using an amplifier based gain stage

When the ADC is measuring a dynamic range of 1V, the lowest resolved voltage a 20-bit ADC would see is 1μV. When a gain stage is used to improve the range, the gain stage also amplifies the noise and makes it prominent enough to affect the ADC’s measurement. This noise affects the number of usable bits that the ADC can provide at this gain setting. Therefore, one has to pick an ADC which gives an optimum resolution at the required gain settings.

The ADCs generally used for measuring the output of a load cell are Delta Sigma (DelSig) ADCs and are redundantly low pass filters. Some DelSig ADCs like the ones found on Cypress’ PSoC3 and PSoC5 devices, have an ability to add the gain in the Delta Sigma modulator itself. The effect of a gain of two on the ADC in this case would result in a change of the ADC’s input range from 0±1.024V to 0±0.512V. Thus, we can achieve higher gains in the ADC’s modulator itself.

There is also a redundancy advantage in doing this. As we increase the gain in the modulator of the ADC, the ADC’s bandwidth starts reducing. This is not of concern for the sensor measurement since the sensor update rates are much smaller. However, the reduction in the bandwidth can be an advantage since it acts as a low pass filter and doesn’t allow noise to enter the system.

Another major concern in a load cell interface is that it is prone to gain error because the output signal range is dependent on the excitation voltage. A small variation in the excitation voltage can cause a similar percentage of gain error in the measurements. We can avoid this if the signal measurements are made as a ratio against the excitation voltage. This can be achieved by two means:

We can measure the signal and excitation voltage separately and then calculate the ratio, thus taking out the gain error. However, this method requires multiplexing of the ADC between the two signals. The other problem is that the signal we are measuring is in the 10mV range and the excitation voltage would be in the volts range. This would mean dynamically changing gain settings and ADC range parameters, which might not be advisable in most analog systems.

The other means of achieving this is by using the reference to the ADC itself. ADCs generally have a reference pin to connect to an external reference. Every measurement made in the ADC is made with respect to the reference. Thus, if we were to provide the excitation voltage or a divided derivative of it as a reference to the ADC, we can achieve a ratio-metric measurement for the signal. Figure 6: Load Cell interface Circuit

Digital filter

We discussed several means for avoiding noise and other error sources in analog signal chains. One of the final stages to get a noise-free output can be achieve using a firmware-based mathematical filter to average out the noise. An easy filter to implement is a moving average filter which uses an array where the input values keep getting streamed in from one side and the oldest values fall of the array from the other side. At any given time, the output of the filter is the average of all elements in the array. Figure 7: Moving average filter

The moving average filter is one of the easiest yet most effective filters to achieve higher noise free bits from your measurement system. The disadvantage is that there is a constant delay which is proportional to the depth of the array being used. That means that change in an n-element moving average filter is going to take n cycles to reflect itself in the output.

This can be a bit misleading if there are larger variations and the output slowly catches up. This condition can be avoided by having a threshold condition check on the variations. If the input varies more than a threshold at a specific point of time, the whole filter is flushed and new data is copied in the filter and also into the output, thus reducing the latency for larger variations.