In hematology analyzers, in vitro diagnostic systems, and many other chemical analysis applications, liquid must be transferred from one container to another in order to remove the sample from the test tube or the reagent from the bottle. These laboratory systems often need to process a large number of samples, so it is important to shorten the processing time as much as possible. In order to improve efficiency, the probe used to aspirate the sample must move quickly, so it is necessary to accurately position the probe relative to the surface of the liquid to be sucked. This article demonstrates a novel method of using capacitance-to-digital converters (CDC), which can be used to complete the work with confidence.
Essentially, the Σ-Δ ADC uses a simple charge balance circuit to apply a reference voltage with a known value and an input voltage with an unknown value to a fixed on-chip input capacitor. The charge balance determines the unknown input voltage. Σ-Δ type CDC is different, and its unknown value is the input capacitance. The known excitation voltage is applied to the input, and the charge balance detects the change of the unknown capacitance, as shown in Figure 1. The CDC will retain the resolution and linearity of the ADC.
Figure 1. Basic CDC architecture
Integrated CDC is deployed in two ways. Single-channel AD7745 and dual-channel AD7746 24-bit CDC work together with a capacitor. One electrode of the capacitor is connected to the excitation output, and the other is connected to the CDC input. Single-electrode devices-such as 24-bit CDCAD7747 with integrated temperature sensor or 16-bit CapTouch™ programmable controller AD7147 can apply excitation to the same electrode and read the capacitance value. The other grounded electrode can be a real electrode or a user’s finger on the touch screen. Both types of CDC can be used for level detection.
In its simplest form, a capacitor can be described as a dielectric material between two parallel plates. The capacitance value changes with changes in the area of the parallel plates, the distance between the two plates and the dielectric constant. Using these variables, you can measure the change in unconventional capacitance and determine the position of the probe relative to the liquid surface.
In this application, the capacitor consists of a conductive plate, which is located under the test tube or moving probe, as shown in Figure 2. The excitation signal is applied to one electrode and the other is connected to the CDC input. No matter which electrode is connected to the excitation signal and which electrode is connected to the CDC input, the measured capacitance is the same. The absolute value of the capacitance depends on the size of the board and the probe, the composition of the dielectric, the distance between the probe and the board, and other environmental factors. Note that the dielectric includes air, test tubes, and liquids in them. This application takes advantage of the changing properties of the mixed dielectric when the probe approaches the board (and more importantly, the liquid surface).
Figure 2. Block diagram of level detection system
Figure 3 shows that the capacitance value increases as the probe approaches the dry test tube. Observation shows that the change is a power series function (quadratic equation), but the coefficient changes with the liquid. Compared to air, liquid has a much higher dielectric constant, so the higher the proportion of liquid in the dielectric, the faster the capacitance rises.
Figure 3. Capacitance measurement of a dry test tube
When the probe is very close to the liquid surface, the measured capacitance value accelerates, as shown in Figure 4. This large change can be used to determine the degree of proximity to the liquid surface.
Figure 4. Capacitance measurement of filled test tube
By normalizing the data, the liquid level can be better determined. If the position of the probe relative to some reference points is accurately known, the system can be characterized in multiple positions without liquid. Once the system has completed the characterization, the data collected during the approach to the liquid surface can be normalized by subtracting the drying data from the approach data, as shown in Figure 5.
Figure 5. Normalized capacitance measurement
In addition to temperature, humidity, and other environmental changes, normalization can also remove the systematic factors of capacitance measurement. The electrode size, the distance between the probe and the board, and the dielectric effect of the air and the test tube do not affect the measurement. At this time, the data represents the effect of adding liquid to the mixed dielectric, making access control more convenient and more consistent.
However, normalized data cannot be used in all situations. For example, the motion control system may not be accurate enough for precise positioning; or the communication link of the motor controller may be slower than the CDC output rate. Even if normalized data is not available, the method described in this article is still feasible.
Use slope and discontinuity
As shown in the figure, as the probe approaches the liquid surface, the measured capacitance accelerates and increases, but this information cannot be conveniently used to control the speed of the probe as it approaches the surface. When the filling level is low, the original capacitance value will be higher than the capacitance value of the container when the filling level is high. With normalized data, the situation is reversed. This adds to the difficulty of finding the threshold-this threshold can be triggered at the right time to change the speed of the probe.
The relationship between slope (or rate of change of capacitance) and position change can be used in the presence of absolute capacitance. When moving the probe at a constant speed, the slope can be approximated by subtracting the previous capacitance reading from the next capacitance reading. As shown in Figure 6, the presentation form of the slope data is consistent with the original capacitance data.
Figure 6. Slope data using normalized capacitance
The slope of the original or normalized capacitance reading is far more consistent under the variable filling level than the reading itself, and it is relatively simpler to find the slope threshold regardless of the filling level. The slope data is slightly noisier than the capacitance data, so averaging it will be useful. When the calculated slope value rises above the noise, the probe is very close to the liquid surface. Using this method, a very stable approach curve can be created.
The data involved so far shows how the system behaves as the probe approaches the liquid surface; but an important feature of this method will become more apparent when the probe comes into contact with the liquid. A large number of discontinuities are generated at this point, as shown in Figure 7. This is not part of the normal acceleration of the capacitance curve as shown by the data point after contact. The capacitance reading at this point is more than twice the reading before the contact. This relationship may change with the system configuration, but it is stable and consistent. The size of the intermittent size makes it relatively easy to find the capacitance threshold, through which the threshold can reliably indicate the degree of breakthrough of the liquid surface. One of the goals of this application is to insert the probe into the liquid a short known distance, so this feature is very important.
Figure 7. Discontinuities on the surface of the liquid
To maximize the throughput rate, the probe should move at the highest possible speed while minimizing the risk of damage caused by the probe being pushed too far. Sometimes a high-precision motor control system may not be provided, so the solution must be able to work without knowing the precise position of the probe. The measurement methods we have discussed so far will allow you to complete this job with confidence.
The flowchart shown in Figure 8 lists the techniques used when approaching liquids.
The probe moves at the highest achievable speed until it is very close to the liquid surface. According to location information, existing computing power, and the ability to pre-characterize the system, this point can be determined by power series calculations, capacitance thresholds, or the slope of the capacitance curve, as described in this article. Averaging the data can more reliably determine this point. Normalizing the capacitance data can also increase the reliability of the system.
When the probe gets close enough to the surface, the speed of the probe drops drastically in order to finally approach the liquid surface. To maximize efficiency, this point should be as close to the surface as possible, but the approach speed must be reduced before penetrating the liquid surface to ensure that the puncture distance is well controlled before the probe stops moving.
The contact with the liquid surface can be determined by the capacitance value and the degree of discontinuity at that point (as described in this article), or by the slope of the capacitance curve. Averaging can reduce noise, but large offsets can be reliably detected without performing this operation. Normalized capacitance data can improve stability, but its impact is not as great as the approach phase.
Subsequently, the probe can be driven to a predetermined distance below the surface. This is easy to do with precision motor control capabilities. If there is no precise motor control, the speed can be estimated, and the probe can move for a fixed period of time.
Figure 8. Simplified control flow chart
After penetrating the liquid, two characteristic data of the capacitance reading will be obtained. First, as the probe moves in the liquid, the change in the measured value is relatively small. Although we expected a constant rate of change to help determine the penetration depth, we did not observe such a phenomenon. Secondly, the measured value changes under different liquid levels are extremely small, as shown in Figure 9. After penetrating the filled test tube and after penetrating the almost empty test tube, the measured capacitance value is basically the same.
Figure 9. The relationship between capacitance and liquid level
However, the normalized data is different. As the liquid level drops, the normalized capacitance value also drops. This feature may be helpful if you want to determine whether the liquid level is decreasing when reliable position data is not available.
How long it takes for the probe to stop after penetrating the liquid surface depends on several factors including the motor control system itself, but a carefully studied approach curve can ensure strict control of the probe and maximize probe speed. In the laboratory, the probe moves approximately 0.45 mm between two capacitance readings at maximum speed, and can stop within 0.25 mm of penetrating the surface. If the sampling rate is higher and the probe moves approximately 0.085 mm between the two samples, it can stop within 0.05 mm of the liquid surface. In either case, the probe works at maximum speed until approximately 1 mm to 3 mm away from the liquid surface to provide the highest efficiency and throughput rate.
This method of breaking the traditional use of an integrated capacitance-to-digital converter provides a simple and stable level detection solution. The proximity curve uses both capacitance and slope measurements to control the movement of the probe. Alternate deployment plans have higher stability or provide more information. This solution can quickly and reliably stop the probe after penetrating the surface, and at the same time move to the final position at the highest probe speed possible. This article only briefly describes the use of CDC technology for level detection. Experienced engineers can use the ideas in this article as a starting point to improve this solution for specific application environments.