Radar is one example of an application that requires pulsed power. In these systems, energy is delivered instantaneously so that minimal voltage droop occurs due to the power demand. Many types of applications fall into this category, from PCMCIA wireless cards to pulsed lasers to radar systems. They are all characterized by a load that has a high instantaneous energy demand on a power system, with limits on the amount of allowable voltage drop.

The load dynamics are defined by the duration of the load pulse, during which voltage must be sustained, and the repetition rate and duty cycle of the pulse. In the case of a pure dc holdup requirement, the pulse length is infinite, and the holdup time will depend on the charge capacity of the capacitor (or battery) supplying the load.

In dynamic loads, there are two sources of voltage droop. The first source is caused by the IR drops along power supply conductors, as well as the equivalent series resistance (ESR) of capacitors used as holdup voltage sources. The second source of voltage droop is voltage decay caused by charge depletion in the holdup device. Thus, for any pulse application, high capacitance can be traded for low ESR as long as the overall holdup requirements are met. However, the ESR and capacitance must be measured at the application frequency, not the component data sheet test frequency. This means capacitor technologies must be evaluated at the anticipated pulse repetition rate, pulse width and pulse current.

The measurement of capacitor performance using these test conditions is shown in Figure 1. As can be seen from the figure:

VTOTAL = I.R(f) + I.t / C(f),

Where I is the pulse-load current, t is the pulse duration, and R(f) and C(f) are the resistance and capacitance, respectively, at the application frequency. According to this equation, the instantaneous voltage drop should be minimized, and for long-term holdup the capacitance should be maximized in order to achieve the necessary voltage holdup for the duration of the pulse. In addition to the pulse-load hold-up requirement, the hold-up device must be fully recharged from the dc power source within the same pulse-load cycle.

For conventional applications, such as PCMCIA wireless cards, low voltages (3.6 V) enable pulsed loads to be driven by pulse supercapacitors or low-profile tantalum chip capacitors. For avionics applications that operate from a 28 V bus, high-voltage solutions are needed, such as tantalum chip banks, wet tantalum, or stacked-ceramic switch-mode capacitors.

In a radar system, the first power stage required to hold up the microwave block must be efficiently filtered from the 28 V bus. When looking for the maximum-capacitance/lowest-ESR options for radar, the switch-mode ceramic solution has the lowest ESR, but also the lowest capacitance and lowest volumetric efficiency. The wet-tantalum solution will have high capacitance, but the ESR of these devices is on the order of Ohms in the 10 kHz range. Tantalum-chip technology has the optimum combination of capacitance and ESR in the milli-Ohm range at 100 kHz, but operation on a 28 V bus requires a 50 V-rated solution as a minimum, so a robust and reliable version beyond the performance of conventional capacitor technology is required.

Low ESR and high reliability

For low-ESR applications, such as airborne radars, a new series recently incorporated into the tantalum MIL-PRF-55365 QPL is the CWR29 family. These include a wide range of CV options rated up to 50 V, with Weibull-reliability grading and optional surge-current testing.

In pulse applications, the high-frequency characteristics and frequency response of a tantalum chip are governed largely by geometry. At higher frequencies, the ESR is dependent upon the overall resistance of the external counter-electrode layer around the tantalum-capacitor element. Tantalum capacitors with a large counter-electrode surface area (provided in packaging with a larger footprint and lower profile) will have lower ESR and less-severe capacitance reduction at higher frequencies.

Because the characteristics of most discrete capacitors are optimized for medium-power applications, it is often necessary to use a bank of capacitors in parallel. For dc buses where the derating rules are conservative, such as the 28 V bus used in avionics, a parallel/series configuration of capacitors can achieve the design targets for capacitance, ESR and voltage derating.

A new series of parts that allows this is AVX's TCP series of stacked-tantalum modules, shown in Figure 2. These modules can be configured in standard two-, four- or six-unit stacks, with other custom configurations available. Specific configuration data is shown in Table 1. The advantage of using such stacks is that the manufacturer can use parts with inset ESR limits and, matching the ESR in a stack, can ensure that current sharing is better equalized. From the design perspective, this advantage gives a large reduction in size, with better packing density on the PC board and greater volumetric efficiency within the system.

MIL-PRF-55365 and reliability management

For mission-critical applications, all CWR series capacitors supplied to MIL-PRF-55365 require a full-lot assessment, which includes voltage conditioning defined as Weibull reliability grading. As tantalum dielectrics cannot be subjected to a high level of overvoltage to accelerate life conditions, the Weibull system uses a combination of voltage and temperature acceleration factors, while the lot is statistically monitored throughout the test (a minimum of 42.5 hours). This grading allows 90% confidence limits, with reliability ranging from 0.1% per 1000 hours to 0.001% per 1000 hours.

The CWR29 series, in addition to having low ESR, also support all the current-availability ratings over and above the availability in CWR06, CWR09 and CWR11. This means, in many cases, the designer also has the option of selecting a higher-voltage rating for a given case-size/capacitance-rating combination, enabling additional voltage derating for the application. Figure 3 illustrates the combined effect of Weibull grading and design derating.

DSCC 07016 — a new reliability standard

The TCP series, based on the CWR29 family, has a long pedigree in aerospace applications. This family has been the mainstay of high-reliability tantalum design since CWR09 was introduced in the 1980s, and are built to the original U.S. MIL-chip footprints adopted for SMD almost a decade earlier.

This is also the only family of molded tantalum chip capacitors to have been kept current, with the CWR19 and CWR29 heralding new case sizes, extended ratings and low-ESR options, which expanded further over time.

By contrast, the sister series of MIL tantalum chip, the CWR11 series based on EIA standard footprints, remains in its original form with no additions after nearly 20 years. This has the advantage of supporting any legacy designs that originally adopted CWR11 (or CWR09, for that matter). The drawback is that if a designer has made an initial prototype with a current EIA standard commercial product, procuring a MIL QPL equivalent for a flight build will be a problem. This is becoming more prevalent as initial design and prototyping is being outsourced.

Suppliers do offer current EIA ratings available with MIL-equivalent screening, which can be catalog items, but sometimes this will still require the designer to create a specification control document (SCD), depending on program requirements.

The new DSCC drawing 07016 provides a complete reference for all the latest ratings and low ESR options available to EIA footprint. However, what makes DSCC 07016 even more important is that it avoids the limitations of previous DSCC drawings. These drawings listed available commercial products, perhaps with some enhanced testing, but certainly not conforming to a high-reliability standard.

However, DSCC 07016 includes full Weibull grading and full surge options to MIL-PRF-55365, and so enables aerospace-grade mission-critical parts to be specified. More important, for pulse applications, stacked modules of DSCC 07016 devices are also available.

Higher-voltage and microwave-power options

So far, the discussion has been limited to 28 V buses used in avionics and the pulse-power loads they supply. To support higher voltages, series combinations of tantalum chip, or stacked ceramic options rated for higher voltages (AVX Hi Voltage SM0), can be used. However, the microwave front-end needs a specialized breed of capacitors. Typical applications include synthetic aperture radar (SARS) and traveling-wave tube (TWT) RF amplifiers.

The latter can benefit from AVX's HQ series power microwave capacitors (Figure 4), which have extremely low ESR at microwave frequencies, giving high-power-handling capability. Their tight tolerance is also ideal for use in matching circuits to increase efficiency.

Providing reliability and redundancy in first-level power

Also available in EIA footprint is the new TBW series of fused tantalum capacitors. These are available to DSCC drawing 04053, but have three significant advantages over earlier fused tantalum types. First, the internal fuse element is made from a thin-film device having the fastest activation speed, the most fail-safe fusing activation, and the lowest intrinsic ESR of all fuses. Second, the internal capacitor can also be wide or graded for high-reliability applications. Third, these fast-acting fused devices can also be supplied in modules for high-density applications.

In a module configuration, by using individually Weibull-graded capacitors, the reliability of the system, even operating in a low-circuit impedance, high-current environment, will be maximized. However, the failure of only a single capacitor can act as a short circuit, pulling down the rail voltage. Having individually fused elements will mean that, for a small amount of capacitance redundancy, line voltages will not be impacted. An additional benefit from this approach is that, with low ESR modules, the current sharing will be better balanced and parallel/series configurations become possible, allowing for use on higher-voltage rails, or for increased reliability through voltage derating.

Emerging pulse-load applications

Airborne radar, as well as an aircraft's electrical and electromechanical control systems, can capitalize on the benefits provided by high-voltage tantalum modules. However, there are more applications emerging where high capacitance at lower voltages are required for high-speed ASIC decoupling. For large systems, the modules described above will provide capacitance on the order of thousands of microfarads, and ESR on the order of tens of milli-Ohms for 4 V to 10 V applications. Yet, for single-chip decoupling, a new technology, the tantalum microchip rated for MIL-PRF-55365/CWR15 reliability specifications, is much more appropriate.

Going back to the characteristics of the counter-electrode in relation to device ESR, element designs with a high surface-area-to-volume ratio will also give the lowest ESR-to-capacitance ratio. This is an important consideration in the design of the CWR15 tantalum microchip, as it has the lowest ESR for micro-miniature tantalum in the 0603 to 1206 equivalent footprint range, and this, combined with low weight, enables increased data-processing capabilities in applications from UAVs to satellites.

Pulse widths of about 600 s, and repetition rates of 500 ms correspond to the transmit pulses for a range of wireless devices, from PCMCIA cards to PDAs and scanners. These are all portable devices typically powered by rechargeable batteries that have too much internal resistance to supply the 2 A to 4 A peak current required. In these cases, where line voltage, as mentioned above, is low (3.6 V to 5.5 V), the use of a low-ESR pulse supercapacitor, such as AVX's Bestcap series, in parallel with the battery will enable the necessary pulse current to be delivered, and the capacitor is then trickle charged from the primary after the pulse. Again, this is an example of combining high capacitance and low ESR to provide instantaneous current with no appreciable voltage droop in the system.

While size reduction for portable systems is a growing design challenge, the major consideration for low-voltage systems remains how to decouple digital logic ICs operating at increasingly high frequencies. The regime for general decoupling for ICs begins at 50 MHz, where capacitance and ESR remain prime considerations. As IC speeds increase beyond this, the designer is actually placing an L-C network — rather than a purely capacitive element — next to the IC to supply the necessary transient switching current. This is done to compensate for the parasitic reactance of the IC at higher frequencies.

Because of this, standard capacitor packaging technology becomes problematic. For example, placing a two-terminal chip on a board creates an inductive loop governed by the geometry of the terminal spacing (and also configuration with the power and ground planes). The smaller the component, the smaller will be the loop inductance, but also the lower the available capacitance for a given dielectric/voltage combination.

There are three methods to overcome this problem. First, the inductive loop for a specific part can be reduced. With AVX LICC series ceramic capacitors, the terminations are on the sides, rather than on the ends of the part, effectively halving the inductance for a given combination of package size, capacitance and voltage rating. The second way to overcome this is to separate the inductive and capacitive properties of the device. AVX's IDC series interdigitated chips are designed to provide internal inductance cancellation when the parts are operated at high-speed ASIC frequencies (specifically, the 50 MHz to 500 MHz range). The third way to overcome loop inductance is to design extremely short signal loops when the capacitor is configured on the PC board. AVX's land grid array (LGA) series use a termination system, coupled with a vertical electrode configuration, to provide extremely small signal paths that provide low inductance at maximum capacitance.

As data rates for these aerospace applications are steadily increasing, these technologies provide significant ASIC decoup-ling capabilities in manned aircraft and unmanned aerial vehicles (UAVs). These capacitor technologies can also be supplied in tin-lead termination and voltage-conditioned versions for use in the most demanding mission-critical platforms.

Table 1. Stacked tantalum modules capacitance and voltage ratings for various cases.
AVX Capacitor Code Capacitance Rating (µF) Rated dc Voltage to 85 °C (V) Package Type (stack layers) ESR (mΩ)
945 9.4 50 2 200
196 18.8 50 4 100
206 20 35 2 200
286 28.2 50 6 67
406 40 35 4 100
606 60 35 6 67
666 66 25 2 85
946 94 20 2 75
137 132 25 4 43
197 188 20 4 38
207 198 25 6 28
207 200 15 2 63
287 282 20 6 25
407 400 15 4 31
447 440 10 2 50
607 600 15 6 21
667 660 6 2 50
887 880 10 4 25
138 1320 6 4 25
138 1320 10 6 17
208 1980 6 6 17


Chris C. Reynolds is an applications manager for AVX Corporation. He has more than 20 years experience with AVX in component manufacturing and applications. Reynolds holds a BSc in physics from Birmingham University, UK and may be contacted by phone at: (843) 444-2868; fax: (843) 626-3123; or e-mail at creynolds@avxus.com.