The third part of our series to explain “ideal diodes” in the small-signal domain and to contrast them with those found in the power domain.
The intent of this series is to illustrate the “ideal” diode in the small-signal domain and contrast it with the ideal diode controllers found in the power domain. In the small-signal domain the descriptor “precision” is favored over “ideal diode.” In the power domain “ideal diode” is unabashedly found throughout—notably applied to special-purpose power management integrated circuits called Ideal Diode Controllers.
This series is provided in three parts:
- Part One: The Ideal Diode in the Small-Signal Domain
- Part Two: Precision Diode Applications in the Small-Signal Domain
- Part Three: Ideal Diode Controllers and Applications in the Power Domain
Ideal Diode Controllers
In the power domain high efficiency is critical. Power losses must be minimized. Many manufacturers have adopted a strategy that uses a power MOSFET as a diode and controls the voltage drop across it. The integrated circuit that serves this function has been dubbed an Ideal Diode Controller. A Schottky diode and an ideal diode controller are contrasted in Figure 18.
The ideal diode controller holds the source-to-drain voltage at 25 mV. This small voltage drop means the attendant power dissipated in the power MOSFET is also very small. For the same circuit conditions, the Schottky diode dissipates 1.26 W, while the MOSFET “diode” only dissipates 75 mW.
Power losses in the diode not only lower efficiency but can also raise thermal management concerns. A power dissipation of only 75 mW will not require a great deal of heat sinking and holds the promise of a small overall system size.
Notice the current from the input voltage source VIN flows from the E-MOSFET source region to the drain region to provide the load current. The n-channel MOSFET is fine with this. The source terminal is also positive relative to the drain terminal. This means the drain current is -ID and we have -VDS. Operation is in the third quadrant of the V-I drain characteristics. The n-channel MOSFET will not be annoyed in the slightest. This is illustrated by the E-MOSFET drain characteristic curves in Figure 19.
Now we are ready to look at the operation of a straight-forward ideal diode controller. Specifically, we examine the Linear Technology LTC4357. It is described as a “Positive High-Voltage Ideal Diode Controller.”
The Linear Technology LTC4357 Ideal Diode Controller
A partial datasheet is provided in Figure 20 listing the device’s description, features, and applications.
Key in its features, the LTC4357 reduces power dissipation by replacing a power Schottky diode with an n-channel E-MOSFET and controlling the voltage drop across it to be only 25 mV. The primary applications include forming a diode OR as required in redundant power supplies and high-availability systems. In the description of the LTC4357, we see “When used in diode-OR and high-current diode applications, the LTC4357 reduces power consumption, heat dissipation, voltage loss and PC board area.” To better understand the LTC4357, we should examine how it works.
How the LTC4357 Works
A functional block diagram is on the device datasheet and is given in Figure 21.
The drawing employs squares to represent the five functional pins. The pins are IN, GATE, OUT, VDD, and GND. The border around the outside represents the outline of the integrated circuit. Unfortunately, the border line has the same weight as the connecting lines, which may provide some initial confusion. The left side provides protection features like the FPD COMP, which is the Fast Pull-Down Comparator. The right side is the “magic” used to maintain 25 mV across the power transistor. To gain further insight, a Multisim simulation was created (see Figure 22). Note: using Multisim in this way is for instructional purposes only. SPICE models exist for LTC4357 for a “proper” simulation.
VIN represents the input voltage source and has been set to 20 V arbitrarily. Using an on-page connector, VIN is also attached to the negative power pin of the LM324 op amp. A 40 V voltage source Vcpump (charge pump voltage source) is connected to the op amp’s positive power pin. Vcpump represents the charge pump voltage, which we will explain shortly. VOFFSET represents the internally-generated 25 mV voltage reference. It was set to 250 mV because the LM324 is not a precision op amp and sports several millivolts of input offset. RLOAD represents the external load connected across the output. The green text indicates the device pin connections IN, GATE, VDD, OUT, and GND. Q1 is an n-channel E-MOSFET that includes a body diode.
The simulation was run and some of the results are given in Figure 23. A multimeter (XMM1) is connected source-drain of power transistor Q1. It indicates the expected 250 mV. A second multimeter (XMM2) is connected from the output to ground. It indicates 19.75 V. Again, this is the expected value. The input is 20 V and the voltage drop across the E-MOSFET is 0.25 V, which leaves 19.75 V for the load.
From the simulation shown in Figure 24, the MOSFET gate-to-source is 5.055 V. This is the enhancement bias required for the MOSFET to supply the load current (19.75 V/ 50 Ohms = 395 mA). Since the MOSFET has its source terminal connected to the 20 V input voltage, the op amp output most be more positive than the 20 V input voltage. In this instance the op amp output voltage is 25.055 V. This requires a charge pump. Basically, a charge pump uses capacitive energy storage to multiply (e.g., double) a voltage. In this case the voltage to be doubled is VIN, and its 20 V value becomes 40 V. This will be reviewed shortly.
Figure 24 provides the last part of the operation puzzle. Multimeter XMM3 displays the voltage at the inverting input of the op amp to be 20 V. By inspection, we see the non-inverting input is connected to VIN, which is also 20 V. This means the op amp’s differential input voltage is 0 V as we would expect.
Basic Charge Pump Operation
The operation of the charge pump is detailed in Figure 25. The circuit is described as a “charge pump” since charge from capacitor C1 is moved to capacitor C2. MOSFET switches are used, and an internal clock is used to toggle the switches. The circuit is also described as a “voltage doubler using switched capacitor technology.” In Figure 25(a) we see state 1. Capacitor C1 is placed across the input voltage source VIN. The capacitor will charge to VIN. In state 2 (Figure 25(b)) capacitor C1 is switched to be in series with the voltage source VIN. Voltage sources in series add. The series combination is connected across capacitor C2, so capacitor C2 will charge to 2VIN. The current draw of the Gate Amplifier is only a few microamps, so capacitor C2 does not discharge appreciably. Therefore, it acts like a solid (stiff) supply voltage.
LTC4357 Protective Functions
As mentioned previously, the left side of the manufacturer’s block diagram in Figure 21 shows protective functions. The block diagram of the protective functions is detailed in Figure 26. The E-MOSFET body diode is shown. When the circuit is operating normally, there is only 25 mV of forward bias across the silicon body diode. The diode will be non-conducting, so the E-MOSFET controls the current. The output voltage across the load will be:
VIN – 25 mV = 19.975 V
The non-inverting input to the Fast Pull Down (FPD) comparator is connected to the output (OUT). Consequently, it will also be at a potential of 19.975 V. The inverting input of the comparator is connected to a 25 mV reference which adds to VIN. Therefore, the inverting input terminal will be at 20.025 V. These results are identified in Figure 26.
In the case of a comparator, when its inverting input is more positive than its non-inverting input, the output of the comparator will be LO. This means the bipolar junction transistor is OFF. This is the normal operating condition.
If the input voltage becomes reduced, we have the situation shown in Figure 27. The comparator now has its non-inverting input more positive than its inverting input. The output of the comparator will be HI. This means the bipolar junction transistor is ON.
The E-MOSFET’s gate terminal is shorted to its source terminal by the bipolar junction transistor. This makes the gate-to-source voltage equal to 0 V. The E-MOSFET goes OFF and the output is effectively disconnected from the input voltage source.
When the output voltage is larger than the input voltage, a reverse current will flow that can damage the input voltage source. The LTC4357 provides fast-acting protection. Reverse current transients are rejected. The 27 V zener diode prohibits damage to the gate-to-source of the E-MOSFET by limiting its maximum voltage.
This video by Rainer Schuster takes you into the lab to look at diode losses and a test circuit for the LTC4357 Ideal Diode Controller. It is in German with English subtitles.
Application of the LTC4357
There are three primary applications for ideal diodes:
- Diode OR for high-reliability redundant power supplies.
- Diode OR selection circuit between the on-board battery and a wall charger.
- Simple reverse-polarity protection for battery-operated devices. The diode is placed in series with the device.
A diode OR configuration is given in Figure 28.
The larger of the two input voltage sources will tend to dominate. If a source voltage droops, the opposite source will pick up the load. The “bare bones” application circuit suggested by the manufacturer can be enhanced by adding additional components (like input and output capacitors). An improvement over the LTC4357 is the LTC4359. Our work thus far has prepared us to quickly grasp its salient points.
Application of the LTC4359
The LTC4359 offers features and capabilities that surpass those of the LTC4357. A comparison is provided in Figure 29.
Unlike the LTC4357, the LTC4359 provides reverse input protection, a shutdown input which enables to serve as a load control, operates over a wider voltage range, and draws a much smaller supply current. Figure 30 shows the block diagram of the LTC4359.
Let’s walk through the block diagram. Two internal MOSFETs connect between the source and gate terminals of the external E-MOSFET. When either of the internal MOSFETs conduct, the gate-to-source voltage of the external MOSFET will be set to zero and it will turn OFF. The shutdown provides load control. In Figure 30 we have a negative comparator that handles reverse input polarity protection, and its output is in an OR connection with the Fast Pull Down (FPD) comparator. Observe the reference voltage has been increased from 25 mV to 30 mV. The gate amplifier is unchanged. However, we note the clock frequency of the charge pump is given to be 500 kHz.
An application circuit is illustrated in Figure 31. Diodes D1 and D2 provide input transient protection. Resistor R1 limits the current through the diodes when they are in conduction. In addition to the “main” MOSFET, a second external MOSFET has been added. For load control, even if the main MOSFET is turned OFF, its body diode will conduct. Therefore, the second MOSFET is reversed and placed in series to eliminate that problem.
This concludes our three-part series illustrating “ideal” diodes found in the small-signal and power domains. You can read the preceding two parts here:
