MENU

Supercaps can be a good choice over batteries for backup applications

Supercaps can be a good choice over batteries for backup applications

Technology News |
By eeNews Europe



Background
The niche that supercapacitors (also known as ultra-capacitors or supercaps) have served in the market space between conventional capacitors and batteries continues to grow as more new applications are found. Supercaps are replacing batteries in data storage applications, requiring medium-to-high current/short duration backup power and battery backup due to sudden disconnect. Specific applications include 3.3V memory backup solid state drives (SSDs), battery-powered industrial and medical portable equipment, industrial alarms and smart power meters.

Compared to batteries, supercaps provide better power density with higher peak power delivery capability, smaller form factors, higher charge cycle life over a wider operating temperature range, and lower ESR. Compared to standard ceramic, tantalum or electrolytic capacitors, supercaps offer higher energy density in a similar form factor and weight. A supercap’s lifetime is maximized by reducing the capacitor’s top-off voltage and avoiding high temperatures (>50°C). See the Table 1 below for a comparison of key features.


Table 1: Supercapacitors vs. capacitors vs. batteries

Summary- Supercaps vs. batteries:
Batteries:
•     Good energy density
•     Moderate power density
•     High equivalent series resistance (ESR) at cold temperatures

Supercaps:
•     Moderate energy density
•     Good power density
•     Low ESR – even at cold temperatures ( ~2x increase at -20°C vs. 25°C)

Supercap limitations:
•     Limited to 2.5V or 2.75V maximum per cell
•     Must compensate for leakage differences in stacked applications
•     Lifetime degrades quicker at high charge voltages and high temperatures

Early generation 2-cell supercap chargers were designed for low current charging from 3.3V, 3xAA, or a Li-Ion/Polymer battery. However, supercap technology improvements have expanded the market, resulting in a slew of medium to higher current opportunities not necessarily confined to the consumer product space. Primary applications include solid state disk drives and mass storage backup systems, high current portable electronic devices such as industrial PDAs and handy terminals, data loggers, instruments, medical equipment, and miscellaneous “dying gasp” industrial applications such as security devices and alarm systems. Other consumer applications include those with high power bursts including LED flash in cameras, PCMCIA card and GPRS/GSM transceivers, and Hard Disk Drives (HDDs) in portable devices.

Design challenges for supercap chargers
Supercaps have many advantages; however, when used in a stack of two or more caps in series, they present the designer with problems such as cell balancing, cell overvoltage damage while charging, excessive current draw and a large footprint/solution. Higher charging current may be required if frequent bursts of high peak power are needed. In addition, many of the charging sources may be current limited, for example, in a battery buffer application or in a USB / PCCARD environment. Being able to deal with these conditions is crucial for space-constrained, higher power portable electronic devices.

Reverse conduction through an IC typically results in a catastrophic event. External fixes such as series rectifying diodes are not very efficient due to high voltage drop. Schottky diodes have less forward drop and therefore higher system efficiency but are more costly than regular diodes. On the other hand, field effect transistors (FETs) offer low on resistances and minimal loss. An internal FET-controlling PowerPath circuit is an elegant way to solve the problem thereby eliminating the potentially damaging consequences. With PowerPath control, in the event the input suddenly drops below the output, the IC’s controller quickly turns the internal FET completely off to prevent any reverse conduction from the output back to the input supply.

Cell balancing series-connected supercapacitors ensures that the voltage across each cell is approximately equal; whereas a lack of cell balancing in a supercap may lead to overvoltage damage. For low-current applications, a charge pump with external circuitry with one balancing resistor per cell is an inexpensive solution to the problem; the balancing resistor value will depend primarily on the capacitor leakage currents as explained below. In order to limit the impact of the current drain due to balancing resistors on supercap energy storage, designers can alternatively use a very low current active balance circuit. Another source of cell mismatch is differences in leakage current. Leakage current in the capacitor cells starts off quite high and then decays to lower values over time. But if the leakage is mismatched between series cells, the cells may become over-voltaged upon recharge unless the designer selects balance resistors that provide significantly more load current on each cap than the cap leakage itself. However, balancing resistors burden the application circuit with unwanted components and permanent discharge current. They also provide no overvoltage protection for each cell if mismatched capacitors are charged at high currents.

For low to medium power applications, another inexpensive (but complicated) approach to solving the supercap charging problem involves using a current limited switch plus discretes and external passive components. In this approach, the current limited switch provides the charge current and limiting, while voltage reference and comparator ICs provide the voltage clamping, and finally an op amp (sink/source) with balance resistors enables supercap cell balancing. Nevertheless, the lower the ballast resistor value, the higher the quiescent current and the shorter the battery run time; the obvious benefit being saved cost. However, this solution is very cumbersome to implement and performance is marginal at best.

Any solution to efficiently satisfy the low to medium current supercap charger IC design constraints outlined above would combine a charge pump-based charger for 2 series supercaps with automatic cell balancing and voltage clamping. Linear Technology has developed a simple, yet sophisticated, monolithic supercap charger IC for these applications which does not need an inductor, eliminates the need for balance resistors, provides reverse blocking, has multiple operating modes and also features low quiescent current.

A simple solution
The LTC3226 is the newest offering in Linear Technology’s family of 2-cell supercapacitor chargers. It is an inductorless supercapacitor charger with backup PowerPath controller for Li-Ion or other low-voltage system rails in applications that require short term backup power. The device employs a low-noise dual mode (1x/2x) charge pump architecture with constant input current to charge two supercapacitors in series from a 2.5V to 5.5V input supply to a programmable capacitor charge voltage between 2.5V and 5.3V. Charger input current is resistor-programmable up to 315mA. The device’s automatic cell balancing and voltage clamping features maintain equal voltages on both cells without requiring balancing resistors. This protects each supercapacitor from overvoltage damage that could otherwise be caused by mismatches in cell capacitance or leakage, while minimizing current drain on the capacitors.

The LTC3226 has two modes of operation: normal and backup. Operating mode is determined by a programmable power fail (PFI) comparator. In normal mode (PFI high), power flows from VIN to VOUT through a low loss external FET ideal diode, and the charge pump stays on to top off the supercapacitor stack. In backup mode (PFI low), the charge pump is turned off and the internal LDO is turned on to supply the VOUT load current from the stored supercapacitor charge while the external ideal diode prevents reverse current flow into VIN. Up to 2A of backup current may be provided from the supercapacitor through the internal LDO.

The LTC3226 operates with a very low 55µA quiescent current when the output voltage is in regulation. The basic charging circuit requires few external components and takes up little space; the IC is offered in a tiny 3mm x 3mm QFN package. The device’s high 900kHz operating frequency reduces the size of external components. Internal current limit and thermal shutdown circuitry allows the device to survive a continuous short-circuit from the PROG, VOUT or CPO pins to ground.  Other features include CAP PGOOD and VIN PFO (power fail) outputs as well as a VOUT RST output for system housekeeping.

The LTC3226 is housed in a compact 16-lead, low-profile (0.75mm) 3mm x 3mm QFN package, with operation from -40°C to 125°C.


Figure 1: LTC3226 block diagram/application

To construct a comparable solution to the LTC3226 requires a convoluted combination of multiple ICs: a buck/boost regulator for SCAP charging, a 2A LDO for backup powerpath, a quad comparator and back-to-back FETs for the external “ideal diode” plus monitoring, and an op amp and various discretes for protection shunts and low-current balancing. Alternatively the user could opt for a “cheap” approach that only charges the SCAPs and provides backup control (without 2 comparators and the op amp); however, there will be no charge current limiting, low-current balancing, cap protection or voltage monitoring features. Compared to more expensive discrete solutions, the cheap method would replace the more expensive high value resistor and op amp combination with inexpensive, low-value resistors that consume a lot of quiescent current and provide no overvoltage protection (clamping) of the supercaps.

Power path control and the diode
The LTC3226 contains a diode controller which controls the gate of an external PFET connected between the input, VIN, and the output, VOUT, through the GATE pin. Refer to Figure 2 for details. Under normal operating conditions, this external FET constitutes the main power path from input to output. For very light loads, the controller maintains a 15mV delta across the FET between the input and output voltage. In the event VIN suddenly drops below VOUT, the controller quickly turns the FET completely off to prevent any reverse conduction from VOUT back to the input supply.

Figure 2. LTC3226 Block diagram

Operating modes
The LTC3226 has two modes of operation: normal and backup. If VIN is above an externally programmable PFI threshold voltage, the part is in normal mode in which power flows from VIN to VOUT through the external FET and the internal charge pump stays on to top off the super capacitor stack. If VIN is below this PFI threshold, the part is in backup mode. In this mode, the internal charge pump is turned off, the external FET is turned off and the LDO is turned on to supply the load current from the stored charge. See Figure 3 for details.

Figure 3: LTC3226 Normal-to-backup mode switching transient waveform

Voltage clamp circuitry
The LTC3226 charge pump is equipped with circuitry to limit the voltage across any supercapacitor in the stack to a maximum allowable preset voltage of 2.65V. If the voltage across the top capacitor (VMID-VCPO) ever gets to 2.65V before the CPO pin reaches the target voltage, the charge pump stops charging the top of the stack via the CPO pin, switches to 1x mode and delivers charge directly to the bottom capacitor via the VMID pin until the stack voltage reaches its programmed value. If the voltage across the bottom capacitor reaches 2.65V before the stack gets to its target value, the charge pump continues to deliver charge to the top of the stack via the CPO pin and a shunt regulator turns on to bleed charge off of the bottom capacitor and prevents the VMID pin voltage from rising any further. The shunt regulator is able to shunt the maximum allowable charge current which is approximately 315mA (in 1x mode). In the event both capacitors exceed 2.65V, the charge pump enters sleep mode by turning off most of its circuitry.

Leakage balancing circuitry
The LTC3226 is equipped with an internal leakage balancing amplifier which serves the VMID pin voltage to exactly half of the CPO pin voltage. However, it has limited source (~4.5mA) and sink (~5.5mA) capability. It is designed to handle slight mismatch of the supercapacitors due to leakage currents; not to correct any gross mismatch due to defects. The balancer is only active as long as the input supply voltage is above the PFI threshold. The internal balancer eliminates the need for external balancing resistors.

Table 2 shows a comparison of Linear Technology’s family of supercapacitor chargers.

Table 2: Comparison of Linear Technology supercap chargers
(Click on image to enlarge)

Conclusion
Supercapacitors are now being used in applications where batteries were once the norm. Initial applications were low current, but technology has advanced and supercaps are now found in a variety of medium and high power applications in both consumer and non-consumer segments. Supercaps have many inherent advantages over batteries such as higher peak power delivery, longer cycle life and smaller form factor. However, product designers using supercaps are faced with problems such as cell balancing and potential over-voltage damage to supercap cells. Fortunately, Linear Technology has addressed these needs by continuously adding to its supercap charger IC family. The LTC3226 is a charge pump based supercap charger with seamless PowerPath control featuring automatic cell balancing, voltage clamping, reverse current protection, various operation modes, low current consumption, and up to 2A of backup current. The LTC3226 offers useful features in a small footprint, reducing overall solution size and in turn enabling more compact, simpler designs.

About the author:
Steve Knoth is senior product marketing engineer, Power Products Group, Linear Technology Corp.

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News

Share:

Linked Articles
10s