Why Are Capacitors Called "non Passive" Devices

introduction

Capacitor is a well-known passive device, which can store energy in electric field. It is used in electronic circuits to prevent DC and allow AC to pass through, so as to smooth the output of power supply and stabilize voltage and power flow. Due to the wide range of applications involved, a wise step is to further understand capacitors and why they are called "non passive" devices.

Active and passive components - is engineering really black and white?

Transistors and integrated circuits are considered active components because they use the energy of the power supply to change the signal. At the same time, we call capacitors, resistors, inductors, connectors, and even PC boards (PCBs) as passive because they do not seem to consume power. However, these apparently passive components can and do change the signal in unexpected ways because they all contain parasitic parts. Therefore, in fact, many so-called passive components are not so passive. In this application note, Part 1 of a series of passive devices consisting of three parts, we will explore the active role of capacitors.

Passive capacitor

Can be defined as inert and / or inert. However, passive electronic components may become active parts of circuits in unexpected ways. Therefore, there is no purely capacitive capacitor at all. All capacitors inherently have parasitic components (Fig. 1).

capacitor © And its largest parasitic element.

Let's take a closer look at the active parasitic elements in Figure 1 below. The capacitor marked "C" is one thing we want to see. All other components are harmful parasites. The parallel resistance RL will cause DC leakage, which will change the bias voltage of the active circuit, destroy the Q factor in the filter, and destroy the holding capacity of the sample and hold circuit. The equivalent series resistance (ESR) reduces the ability of the capacitor to reduce ripple and transmit high-frequency signals, because the equivalent series inductance (ESL) produces a tuning circuit (i.e. a circuit with self resonance). This means that above the self resonant frequency, the capacitor appears to be inductive and can no longer couple high-frequency noise from the power supply to the ground. The dielectric can be piezoelectric, which increases the noise generated by vibration (AC) and looks like a battery inside a C capacitor (not shown). The piezoelectric effect caused by cooling solder stress will change the value of the capacitor. Polarized electrolytic capacitors can also have parasitic diodes (not shown) in series, which can rectify high-frequency signals and change bias or increase unnecessary distortion.

Small cells SB1 to Sb4 represent Seebeck junctions where different metals (parasitic thermocouples) establish voltage sources. When connecting test equipment, we need to consider the Seebeck effect of ordinary connectors. Appendix J in Jim Williams' application note, figure J5 shows that the thermoelectric potential of BNC and banana connector pairs ranges from 0.07 µ V / ° C to 1.7 µ V / ° C. This difference is just a simple connection we make in the laboratory every day. Multiplying the seemingly small offset gain by 1000, we get a voltage of 1.7mv, which is before we actually carry out any production.

SB2 and Sb3 can be inside a capacitor in which the foil is connected to the lead, or the metallization layer is connected to the coating or solder in the surface mounted part. SB1 and Sb4 represent the connection points from the part to the PCB copper trace through solder. Solder used to be simple 63% lead and 37% tin. But today, people have to ask about the alloy content, because lead-free RoHS solder will change greatly and affect the voltage around the capacitor.

The so-called "immersion" of dielectric absorption Da or Bob pease can be modeled as countless different RC time constants, DA1 to dainfinity. These time constants are composed of resistance RDA and capacitor CDA. Bob pease provided us with some practical examples of "soaking" time. I remember the interesting experience of soaking in the appendix.

"Well, if you turn off the color TV and open the back cover, what is the first thing you need to do before starting the operation? Put the grounding strap on the screwdriver and extend it under the rubber sleeve on the HV plug to release the CRT. Well, since the capacitor has been discharged, how much voltage will it" absorb "back to the" capacitor "of the picture tube if it is allowed to stand for about 10 minutes In the second discharge, it is enough to form a visible arc... Now this is what I call dielectric absorption. "

Therefore, the capacitor can change the capacitance with the applied voltage. Then, coupled with typical aging, temperature dependence and various ways in which capacitors may be physically damaged, this simple passive component becomes more complex.

Now, we should talk about self resonance, which is the most common capacitor problem of decoupling capacitor and poor grounding. If the grounding is poor, the capacitor will not work. The capacitance self resonance is affected by the ESL shown in Figure 1. However, do not ignore the influence of PCB vias. At RF, these vias will affect the self resonance point of small capacitors. Check Figure 2 and focus on the 1 µ f curve.

Self resonance of three capacitors (the lowest point on the curve). The chart shows that the performance of capacitors is not exactly the same. On the left side where the routing (impedance) moves down, the capacitor acts as a capacitor. However, when they reach the lowest point and start up, they become inductors (ESL) and are no longer effectively used as decoupling capacitors.

The minimum value of 1 µ f routing is found at 4.6mhz. Above this frequency, ESL is dominant, and the working mode of capacitor is similar to that of inductor. This tells us that the decoupling capacitor is a bidirectional pipeline for high frequency: the high frequency on the power bus is shared with the ground, and vice versa. Capacitors equalize the difference between power and ground.

Considering more information about signal frequency and capacitor, we may forget the generated harmonics or sidebands. For example, the actual 50MHz square wave SPI clock will have odd harmonics to infinity. Most systems, but not all systems, can ignore harmonics higher than the fifth harmonic because the energy is so low that it is lower than the background noise. However, if the harmonic is rectified in the semiconductor and can still be converted into new low-frequency interference, it will still cause problems.

Manipulating manufacturing tolerances

Figure 2 shows that not all capacitors are equal. Generally, high-quality capacitors have high repeatability, and some cheap capacitors can exchange large manufacturing tolerances for lower cost. Some manufacturers "pack" (Figure 3) or choose capacitors with strict tolerance, which will be sold at a high price. If the capacitor is used to set the time or frequency in the system, it may be harmful.

Combining or classifying manufacturing tolerances can affect capacitor performance in different ways.

The solid line (black) curve in Figure 3 is the standard deviation of a good manufacturing process. Although we used this illustration for resistors in Maxim integrated's application note 4301 "zero transistor IC, a new platform for IC Design", the data also applies to capacitors. As manufacturing tolerances change, the number of parts in each bin will also change. The tolerance may move to the right (green dotted line), resulting in no yield under the 1% tolerance. It can be bimodal (gray dotted line) with many 5% and 10% tolerance parts and only 1% and 2% tolerance parts.

Box "seems" to ensure that the 2% tolerance range is only from negative 1 to negative 2 and from 1 to positive 2 (i.e. there is no 1% part). It also "seems" to delete any parts with 1% and 2% tolerances from the 5% packing. We say "seems" and "appears" because sales and humanity also affect the combination of the two. For example, the factory manager may need to ship capacitors with a tolerance of 5%, but his demand is insufficient this month. However, his tolerance is indeed 2%. So this month he threw them into the 5% dustbin and shipped them. Clearly, human intervention can and does distort statistics and methods.

What does this mean for our passive capacitors? We must understand that we may expect tolerances, such as ± 5%, and there may be ± 2% holes in the middle. If the capacitor controls the critical frequency or timing, we need to take this into account. This may also mean that we need to plan to correct for larger changes through calibration.

How does welding affect passive performance

Welding introduces stress in capacitors, especially in surface mounted parts. This stress will cause the piezoelectric voltage to vibrate and even rupture the capacitor, resulting in the subsequent failure of the capacitor.

It is impressive to see the correct reflow soldering. The surface tension of the molten solder rotates the parts into alignment, as if by magic. However, poor solder temperature curve will indeed damage the device. Do you see the capacitor standing at one end like a tombstone? This can happen if the solder temperature rises incorrectly. Always follow the manufacturer's solder profile recommendations. Some components are more sensitive to temperature, so circuit board components may require two or more solders with different melting points. First, most components in the circuit are welded with solder with the highest melting point, and then any "sensitive" components are welded at a lower temperature. The solder must be used in the correct order so that those parts welded earlier in the process will not be welded later.

summary

When we talk about passive devices such as capacitors, we must remember that these devices contain parasitic parts that can change the signal. Of course, the impact depends on the signal strength. If microvolts are to be measured, everything is important: grounding (star point), shielded decoupling capacitors, protection, layout, Seebeck effect, cable construction and welded connectors. Our schematics usually hide this, which is acceptable before we look for small noise or voltage.

Remember that a passive capacitor is just a component and is actually more active than it looks. Component parasitic effects, tolerances, calibration, temperature, aging and even assembly methods and practices will have a slight impact, which will affect the performance of the equipment. Knowing this, we need to understand the potential errors that many capacitors may accumulate. In this three part series of future application notes, we will discuss other so-called passive components: resistors, potentiometers, switches and surprisingly low-level PCBs.

Finally, AVX and KEMET are capacitor companies that specify parasitic components and provide free spice tools. These spice tools enable us to plot the actual performance of capacitors. The application notes on both websites are also very useful.

Editing: hfy