Soul-searching question: Is my voltage reference source design sensitive to humidity?

While every electronic design will make different compromises in terms of performance, the general analog signal chain will have some form of analog input signal conditioning, such as ADC and voltage reference. To help illustrate the main purpose of this article, we will use a medium-speed 100kSPS, 16-bit analog sensor signal input design as an example, as shown in Figure 1. For more information on some of the design trade-offs and design choices for this signal chain, see the CN-0255 circuit note.

Figure 1. 16-bit signal chain functional block diagram.

The 2.5 V voltage reference used in this application is the ADR4525 from the ADR45xx plastic package voltage reference family, which provides high accuracy, low power, low noise, and has ±0.01% (±100ppm) initial accuracy, excellent temperature stability, and low output noise. The low thermally induced output voltage hysteresis and low long-term output voltage drift of the ADR4525 improve system performance. The maximum operating current of 950μA and the low dropout voltage (maximum) of 500mV make this device ideal for portable equipment.

Once you have selected the components to be used in your precision analog signal chain, it is time for the PCB assembly team to produce a repeatable system using the printed circuit board as the substrate for the electronic design. Anyone who has worked with precision electronics knows that board-level mechanical stress manifests itself in the form of DC bias in precision circuit designs or MEMS-based sensor designs. Verification is simple, just press the plastic package of the voltage reference and see a change in the output voltage or sensor output. Environmental factors such as moisture and temperature can affect electronic device performance because moisture/humidity/temperature cause differential stress. Temperature causes mechanical stress in the package and board due to the different coefficients of thermal expansion of the materials used to make the package and board. Moisture causes mechanical stress in the package and board because both the plastic and board absorb moisture and expand.

在塑料封装电压参考中，因为环境原因产生的机械应力往往表现为随温度/时间变化产生漂移，在塑料封装MEMS加速度计中，则表现为增加偏移量。对于塑料封装，湿度导致的机械应力相当显著，要控制这种湿度效应，方法之一就是将集成电路封装到陶瓷或密封封装中。虽然此方法能解决大量与湿度有关的挑战，但这种解决方案会额外增加封装成本，且通常会导致元件尺寸更大。

Another way to separate these stresses from the reference voltage is to use conformal coatings during PCB manufacturing so that any mechanical stress on the board will have a smaller effect on the reference voltage. In this case, a thin composite coating is applied to the voltage reference and the corresponding PCB to ensure that stresses caused by moisture or temperature on the PCB are not completely converted into differential stresses on the reference voltage chip package and cause drift. This also ensures that the impact of low-temperature condensation on the package is reduced.

HumiSeal ^® is a specialty coatings manufacturer that offers a wide range of conformal coatings including acrylics, polyurethanes, silicones, epoxies, and water-based coatings for protecting sensitive components in PCB production. The suitability of the selected coating can be determined by the water vapor permeability (MVP) parameter, which is the rate at which water vapor passes through the coating. This is extremely important since we are trying to make the PCB impervious to humidity.

The method for testing MVP is to take dry cups, apply the corresponding coatings, place them in a constant humidity chamber, and then weigh the cups regularly to assess how much water passes through the coating into the dry cups. The week-long test showed that these coatings can effectively slow down the speed of water passing through.

Table 1 shows the MVP ratings and material thicknesses for various conformal coating options.

Looking at the data in the table, it can be seen that in all cases (except for the very thick UV-cured coating material, UV40), these coatings will experience some degree of water penetration over time. This is measured based on the weight of water that seeps through a given surface area of ​​the coating over a given period of time; in these measurements, the time period was seven days. Selecting the commonly used 1A33 coating, a convenient, cost-effective polyurethane coating, the results show that it is more than 10 times more effective at slowing the rate of water vapor absorption than a rubber-based 1B51 coating of the same thickness. However, the important conclusion to draw from this table is that these coatings cannot completely block water penetration after being placed in a high humidity environment for a long enough period of time.

This does not negate the use of conformal coatings. Rather, it helps to understand the environment in which the electronic device will be exposed. Will the exposed electronic device only experience high water vapor permeation for short periods of time? Will the packaging/container of the electronic device block water vapor permeation? Is conformal coating as useful as wearing a belt and suspenders at the same time? The environment in which the electronic device is exposed changes so frequently, is conformal coating used only to make the electronic device perform better in environments that change too quickly? It is important for product owners to understand all of these issues before starting to use conformal coatings.

Before we get into the actual data, one thing to consider is that the use of conformal coatings can increase mechanical stress in some cases. This is because the coating can increase package stress if not applied correctly. For example, during the PCB manufacturing stage, if the surface of the voltage reference package contained moisture before coating, it is almost certain that this moisture will penetrate into the hydrophilic plastic package. From the data sheet for the 1A33 product: "The cleanliness of the substrate itself is critical to the successful application of a conformal coating. The substrate surface must be free of moisture, dirt, wax, grease, flux residues, and all other contaminants. Contaminants beneath the coating can cause problems that may result in assembly failure." This is important to note for anyone considering the use of conformal coatings.

To evaluate the effects of conformal coating, ADI created a set of test boards. Each test board had 27 identical high performance voltage references soldered to the PCB using the recommended J-STD-020 reflow method. The boards were placed in a humidity chamber and measured using a Keysight 3458A 8.5-digit DMM (Model 002) and verified to achieve 4 ppm/year drift using the LTZ1000. The humidity chamber maintains a constant temperature and humidity to allow the boards to stabilize. The boards were placed in the humidity chamber for one week, after which the temperature was maintained constant and the humidity was increased. Two different conformal coating processes were used on the plastic encapsulated voltage references to evaluate the effect of humidity on the coating.

Figure 2. ADR4525 voltage reference in a ceramic package.

Using the ceramic packaged ADR4525 as a reference (Figure 2), the output voltage changed by about 3ppm, or 0.075ppm/%RH, after 100 hours in a 70% humidity environment, indicating excellent stability for the ceramic package. The first peak in the data is due to the temperature jump caused by the sudden change in humidity. The temperature in the humidity chamber slowly warmed back up to 25°C, as can be seen in the data. In contrast, when the plastic packaged voltage reference chip was placed in the same environment and test conditions, its voltage output changed by about 150ppm, as shown in Figure 3. Normalizing the data in Figure 3 for a 60%RH drift shows that the output drift was about 2.5ppm/%RH without the conformal coating. In addition, it is clear that the drift did not completely stop after the board was placed in the high humidity environment for 168 hours.

Figure 3. The ADR4525 reference voltage in a plastic package is affected by humidity levels between 20% and 80%.

Next, the HumiSeal 1B73 acrylic coating was tested, and the data is shown in Figure 4. The application steps are as follows: the board is first cleaned and dried (quickly dipped into 75% isopropyl alcohol and 25% deionized water several times, lightly brushed by hand, and then baked at 150°F for 2 hours), and then sprayed with the specified thickness of 1B73 coating. The entire board is coated except for the edge connectors, and the board must be clean before the output voltage can be measured.

Figure 4. HumiSeal 1B73 acrylic coating is spray-coated on the ADR45xx voltage reference.

The oven used in this experiment was limited to 70%RH with a normalized drift of about 100ppm/40%RH or 2.5ppm/%RH, not much different than when no coating was used. After consulting with HumiSeal, it was learned that the coating may not have fully blended to the bottom surface of the voltage reference package and the edge of the device. It is also important to note here that the 168 hours of testing in a high humidity environment may not have been long enough, as the voltage reference did not appear to have fully stabilized, similar to the uncoated device. However, it is worth noting that the rate of change of the humidity effect seems to have slowed, at least initially, which provides support for the concept of moisture permeability, that the coating does not prevent moisture, but slows it down.

The next test attempt used the same conformal coating (HumiSeal 1B73) but in a deep-dip, three-step coating process to ensure complete coverage of the entire board. The data is shown in Figure 5.

Figure 5. ADR45xx voltage reference is coated with HumiSeal 1B73 acrylic coating using a deep-dip, three-step coating process.

Due to oven issues, this test could not exceed 96 hours. The normalization step for the data from 30%RH to 70%RH showed a drift of around 90ppm or 2.3ppm/%RH, which did not achieve the dramatic improvement that this application process was intended to achieve, but there was a slight improvement in the spray coating, which of course could be said to disappear if the test time was longer. Table 2 summarizes the three tests.

Table 2. Summary of Humidity Testing Using Conformal Coating

Future testing may include other types of conformal coatings (silicone, rubber, etc.) and many variations in the application process. Additionally, cross-sectional analysis after coating can confirm that the applied coating thickness meets the manufacturer's requirements and that certain edge locations have adequate coating. Overall, these experimental data indicate that ceramic hermetic packaging is the only ideal defense against moisture ingress.