People have been predicting the demise of low-drop-out linear regulators in cars for many years. Yet low-drop-out (LDO) regulators continue to survive and even thrive because they're cheap and easy to use. In this article I'll highlight the complexity of LDO regulators, look at what's new in the market (yes, there are some advances), and examine the move toward switching regulators as automotive power requirements continue to rise.
New linear regulators
What's new with linear regulators? Let's start with the output capacitors. Ceramic capacitors in the familiar 0402 package are the capacitors of choice these days. That's mostly because improved materials have raised their temperature range from 125°C (257°F) to 150°C (302°F) and improved mounting methods reduce thermal shock and improve vibration resistance. These caps' small size reduces their inductive component, for better high-frequency performance. But the key trait of concern for ceramic capacitors is their low equivalent series resistance, or ESR.
The basic closed-loop linear regulator system consists of an error amplifier, output driver, and loads. Figures 1 through 3 detail the closed-loop frequency response of a bipolar linear regulator, highlighting the effect of changing the output capacitor ESR while keeping system setup and output capacitance values the same. A capacitor with a 1 ohm ESR (Figure 1) is stable, while one with a very low 0.01 ohm ESR (Figure 2) is unstable; the larger ESR of 3 ohms (Figure 3) is also unstable.
A rule of thumb for switching power supplies–that whenever the closed-loop gain is greater than or equal to 1, the closed-loop phase will never come to within 30 degrees of 360–is also very applicable here.
Most linear regulators don't give you any accessible points to measure stability curves. Instead, chip manufacturers include graphs that show where the expected areas of stability are versus the output capacitor's ESR value. Figure 4 shows the typical difference in unstable and stable regions for the output capacitor ESR, dependent on the output regulator voltage over a variation in output current. Figure 5 shows the difference in unstable and stable regions based on the output capacitor value.
Load response times will typically vary inversely with IC stability product regions. The loop response time has been slowed down to provide better stability. An external output capacitor will compensate for most transient requirements. Be sure to provide a capacitor large enough for your requirements. Use your typical capacitor equation:
to calculate the capacitor value based on the system load magnitude, transient time, and system allowable drop in output voltage.
Quiescent current has become an important specification, both when a regulator is being loaded and when it's in standby. Historically, quiescent current was not a concern. As electronic content in automobiles has increased, the limitations of present use batteries and alternators have been reached. Semiconductor fabrication processes can have an effect on the amount of quiescent current, an effect that we can see in the typical performance characteristics for products manufactured on two different types of fabrication. Figure 6 shows a device manufactured on a bipolar process while Figure 7 shows a device manufactured on a BCD process. Note the flat line nature of the device fabricated on the BCD process.
The bipolar process results in a device that requires increasing quiescent current at higher loads. The device manufactured on the BCD process will preserve the low quiescent current at low loads when operated at the higher loads. The result is a lower contribution to the module quiescent current limitation.
You can use a watchdog regulator in a current-saving maneuver. Watchdog regulators save current by sending a wakeup signal to a microprocessor. While microprocessor instructions are being set in motion, a concurrent signal is sent back to the voltage regulator by the microprocessor, notifying the regulator that it must maintain regulation. Once the microprocessor completes its commands and directives, the feedback signal back to the regulator is removed. The watchdog regulator recognizes this event and sends a reset signal back to the microprocessor, shutting it down, shown in Figure 8. The end result is less current draw until the microprocessor's efforts are needed again.
Another current-saving scheme new to the IC-regulator world is to momentarily power-down unneeded circuitry. Any part of the regulator that's not immediately needed can be powered down and operated in a pulsed on/off mode.
This scheme works best for lightly loaded conditions at cold or room temperature. Increased leakage current at higher temperatures (reached through an ambient temperature rise or die temperature rise caused by on-chip power) complicates proper operation of the device.
Interest in dual regulators (two independent output regulators on one chip) is increasing. Several microprocessors now require a dual supply voltage. One supply (usually the lower voltage) powers the core and the second powers the I/O. Lowering the core voltage allows more transistors to be squeezed on chip without melting the device or exceeding the thermal limitations of its package.
While dual linear regulator use is not intended as a quiescent current savings tool (it's more for convenience, space savings, and cost), they contribute to the power savings and distribution of power in the system. Current savings is a result of the common use of circuitry (such as the bandgap reference voltage and the current source bias strings) within the dual regulator.
Integrating multiple regulators on a single IC is good for convenience, space savings, and cost but is limited by the amount of allowable power in the IC.
Packaging is another area where improvements have allowed more power to be dissipated in a single package. Improvements in thermal resistance have been achieved by accessing the metal lead frame material (exposed pad). Heat can be dissipated more efficiently through the metal connection compared with its plastic counterpart. Figure 9 shows a typical exposed pad (epad) package. This device is a 300-mil, 16 lead SOW epad package, with an epad dimension of 150 mil x 184 mil.
Exceeding the manufacturer's limits on temperature (usually around 150°C/302°F at the junction) can either damage the regulator instantly or lead to early failures due to stress caused by the different thermal expansion coefficients of the silicon, the wire bonds, and the plastic package. Failure rates go up exponentially as temperature increases. Research is under way to increase the acceptable operating temperatures of these electrical components.
Switching regulators will see increased use when all the new linear regulator options I've described have been used. Switchers are more expensive than linear regulators because of their external component count. They also bear the hidden costs of troubleshooting. By their nature, switchers exhibit properties that require skill and education in electromagnetic interference to tame.
There's no argument that switching regulators are more efficient than LDOs. A switching regulator can exhibit 90% efficiency versus 36% for an LDO as shown in Table 1. Figure 10 displays a typical efficiency curve from a buck switching regulator.
Power loss in a linear regulator is simply the current to the load (ignoring any quiescent current) multiplied by the voltage difference between the input and output. The example in Figure 11 shows a 9V drop across the regulator so the efficiency works out to 35.7%, independent of the load but dependent on input voltage as the next example will illustrate.
The only way to increase the efficiency of the linear regulator is to lower the voltage across it. You can do this by using a switching regulator alongside the linear regulator (shown in Figure 12), creating a 6V DC input to the linear regulator compared with the 14V DC input in the previous example. A switching device can efficiently step down to a more manageable voltage and distribute this voltage to run other linear regulators. This arrangement provides the improved efficiency of a switcher with the cost savings of a linear regulator. Eliminating the connection from the linear regulator to the battery can reduce cost even further because a lower voltage part can be used.
The improved efficiency provides an overall system efficiency of 74.7% as shown in Figure 13, compared with the 35.7% from Figure 11.
Another way to reduce current drain involves combining the function of a switching regulator with a linear regulator. Switching regulators are most efficient when they're driving the loads for which they were designed. When the output voltage is not heavily loaded, however, the current needed to keep the switching regulator switching becomes more of a burden than an attribute. A linear regulator can be more efficient under these conditions.
Figure 14 shows a device capable of switching between modes of operation as a linear regulator and as a switching regulator. The block is divided into two independently operated sections: the first section (green) supports the system for use as a buck switching regulator. The second section (yellow) supports the system for use as a linear regulator. The temperature-independent voltage reference is shared between the two sections. Engineers can choose to change modes based on output loading for best efficiency or EMI requirements. At very light loads, efficiency will be higher for a linear regulator. At heavier loads efficiency will be higher for a switching regulator. The EMI performance of the linear regulator will always be superior to the switching regulator.
Other applications with switching regulators involve starting an automobile. With modern automotive security systems, the key must be verified as belonging to the vehicle it's trying to start. The heavy load from an engine start causes a severe dip in battery voltage, which must not cause a brown-out or reset in the microprocessor involved in the verification.
To accomplish this, you need a device that can provide a boosted voltage as well as a step-down or buck voltage to the system. One way to do this is to use a SEPIC (Single-Ended Primary Inductance Converter) as shown in Figure 15. Capacitor C1 must be able to withstand normal high-voltage operation restrictions (in terms of flyback pulses) of this system and the high-voltage load dump pulses (among other transients) that it sees through inductor L1. This means a high-voltage capacitor and, to maintain the best efficiency, a low ESR value for the capacitor because of the high currents passing through this component. Some engineers aren't comfortable with high currents flowing through capacitors. Overheating can occur resulting in reliability issues or capacitor deterioration (short or open circuit).
Another solution to this problem is to use a pass-through zone technology with a bimodal converter. This creates a smooth transition between the buck and boost modes of operation while creating the required buck/boost operation. This is shown in Figure 16.
In normal operation, transistor Q1 operates as a buck switching regulator while the control circuit holds Q2 off. When the input voltage of Vbat dips down, Q1 turns on 100% while Q2 operates the circuit as a boost switching regulator. The resistor Rpassthrough helps to set up operation through which there is a pass-through zone whereby the regulated output voltage is slightly altered to provide a smooth transition when toggling modes of operation.
Available electrical power is limited in the present automobile system. From an engineering perspective, the switching regulator offers the best efficiency over the linear regulator. The switching regulator will enable the most integrated features but will also add the most cost to the system. Will the automotive consumer pay for the most features (switchers) or settle for lesser features (linear regulators)?
Frank Kolanko is a principal applications engineer at ON Semiconductor. You can reach him at .
Motoki, Tomoo. “Chip Monolithic Ceramic Capacitors Drive Automotive Applications,” AEI, Dempa Publications, Inc. August 2003.
Brown, Marty. Practical Switching Power Supply Design . Motorola Semiconductor, ISBN 0-12-137030-5, Academic Press, March 1990.
ON Semiconductor documents:
O'Malley, Kieran, “Compensation for Linear Regulators,” ON Semiconductor SR003AN/D application note, April 2001: www.onsemi.com/pub/Collateral/SR003AN-D.PDF
“NCV8503 Series Micropower 400 mA LDO Linear Regulators with ENABLE, DELAY, Adjustable RESET, and General Use Comparator,” February 2006: www.onsemi.com/pub/Collateral/NCV8503-D.PDF
“NCV8508 5.0 V, 250 mA LDO with Watchdog and RESET,” November 2005: www.onsemi.com/pub/Collateral/NCV8508-D.PDF
“NCV8800 Series Synchronous Buck Regulator with 1.0 Amp Switch” March 2006: www.onsemi.com/pub/Collateral/NCV8800-D.PDF
“NCP1500: Dual Mode PWM/Linear BUCK converter,” February 2004: www.onsemi.com/pub/Collateral/NCP1500-D.PDF
“SOIC 16 Lead Wide Body, Exposed Pad Case 751R 02 Issue B Mechanical Case Outline Package Dimensions,” July 2005: www.onsemi.com/pub/Collateral/751R-02.PDF