Abstract: An LED driver is disclosed that boosts an input voltage to drive any number of LEDs in series. The driver includes a switch-mode current regulator that supplies regulated current pulses to the LEDs. No high voltage output capacitor is used to smooth the current pulses, so the LEDs are turned on any off at the switching frequency. Also, no blocking diode between the switching transistor and the LEDs is used. The cathode of the “bottom” LED in the string is connected to ground via a current sense resistor. In parallel with the sense resistor is connected an RC filter using a small, low voltage filter capacitor. The RC filter provides a substantially smooth feedback voltage for the current regulator to control the duty cycle of the switching transistor so that the feedback voltage matches a reference voltage.

Abstract: In a buck-boost converter, the method compensates for the boost mode power switch having a minimum on-time when entering the buck-boost mode from the buck mode by immediately decreasing a duty cycle of the buck mode power switch upon entering the buck-boost mode. This prevents the inductor current from being higher at the end of the switching cycle than at the beginning of the cycle, so the output voltage stays regulated without the converter oscillating between the buck mode and the buck-boost mode. The duty cycle of the buck power switch is increased in the buck-boost mode as the input voltage further falls and the boost power switch duty cycle is increased. Upon transitioning into the boost mode, the duty cycle of the boost power switch is immediately reduced to compensate for the buck switching being stopped and the buck power switch having a minimum off-time.

Abstract: A method of forming a lateral DMOS transistor includes performing a low energy implantation using a first dopant type and being applied to the entire device area. The dopants of the low energy implantation are blocked by the conductive gate. The method further includes performing a high energy implantation using a third dopant type and being applied to the entire device area. The dopants of the high energy implantation penetrate the conductive gate and is introduced into the entire device active area including underneath the conductive gate. After annealing, a double-diffused lightly doped drain (DLDD) region is formed from the high and low energy implantations and is used as a drift region of the lateral DMOS transistor. The DLDD region overlaps with the body region at a channel region and interacts with the dopants of the body region to adjust a threshold voltage of the lateral DMOS transistor.

Abstract: A MOS transistor includes a body region of a first conductivity type, a conductive gate and a first dielectric layer, a source region of a second conductivity type formed in the body region, a heavily doped source contact diffusion region formed in the source region, a lightly doped drain region of the second conductivity type formed in the body region where the lightly doped drain region is a drift region of the MOS transistor, a heavily doped drain contact diffusion region of the second conductivity type formed in the lightly doped drain region; and an insulating trench formed in the lightly doped drain region adjacent the drain contact diffusion region. The insulating trench blocks a surface current path in the drift region thereby forming vertical current paths in the drift region around the bottom surface of the trench.

Abstract: An overcurrent protection circuit for a current setting circuit is disclosed herein that prevents a user-selectable current from exceeding a current limit when an incorrect current selecting component (or current selecting circuit) is connected to an external control pin of a package by the user, or when the control pin is inadvertently grounded. The protection circuit senses a current (A1*Iset) mirrored from the user-set current (Iset). If the mirrored current is above a threshold, the protection circuit limits the Iset current to be at or below a current limit level. In one embodiment, the protection circuit comprises a transistor that turns on when the mirrored current exceeds a threshold, and the transistor shunts control current from a series transistor generating the user-set current Iset.

Abstract: Various circuits are described herein where a series transistor used to control current through a string of LEDs, driven by a high voltage, is not subjected to the high voltage when the transistor is turned off pursuant to a PWM signal. To avoid the transistor experiencing the high voltage, the HV regulator is disabled shortly before the transistor is turned off and is enabled shortly after the transistor has turned back on. Control circuits for controlling the regulator and transistor include delay circuits and/or voltage sensing circuits to ensure that the transistor is always on prior to the voltage regulator being enabled pursuant to the incoming PWM signal, and the voltage regulator is always disabled when the first transistor is off pursuant to the incoming PWM signal.

Abstract: A power budget monitoring circuit in a multi-port PSE includes a differential amplifier and a transistor for setting a reference voltage across a first resistor to establish a reference current, multiple current mirror output devices each associated with a power port of the PSE, a second resistor and a comparator. Each current mirror output device provides an output current indicative of the power demanded by the associated power port where the output currents are summed at a second node into a monitor current. The second resistor has a resistance value proportional to a maximum power budget of the PSE and receives the monitor current. A monitor voltage develops across the second resistor indicative of the total power demanded by the power ports. The comparator compares the monitor voltage to the reference voltage and provides a comparator output signal indicating whether the maximum power budget of the PSE has been exceeded.

Abstract: A voltage regulation system is provided including detecting a feedback voltage less than a reference voltage; asserting a current source gate output by the feedback voltage less than the reference voltage; activating a gated current source by the current source gate output; and waiting a delay interval before negating the current source gate output for turning off the gated current source.

Abstract: In an average-current mode control type buck-boost PWM converter, a sample and hold circuit is inserted in the current loop to avoid problems associated with ripple of the average inductor current demand signal. The rippling average inductor current is generated by a differential transconductance amplifier having applied to its inputs an error signal and a signal corresponding to the instantaneous current through the inductor, where the output of the amplifier is filtered. The rippling average inductor current is sampled and held at the beginning of each switching cycle, prior to the average inductor current demand signal being compared to buck and boost sawtooth waveforms. By using the sample and hold circuit, the feedback loops are easier to stabilize, and the converter cannot switch modes during a switching cycle.

Abstract: An LED driver is disclosed that drives LEDs connected in parallel. Instead of applying current to all the parallel-connected LEDs at the same time, under control of a common PWM brightness control signal, the application of current to each parallel path is staggered by using staggered brightness control signals. The turning on of the LEDs in the different parallel paths will have the same duty cycle but will be out of phase. This reduces ripple in the power supply by reducing the magnitude of the instantaneous current sink. In one embodiment, a shift register contains a binary representation of the PWM duty cycle, and a clock shifts the bits along the shift register. The PWM brightness control signals for each parallel path of LEDs are tapped from different positions along the shift register so that the PWM brightness control signals are identical but staggered.

Abstract: An Ethernet repeater system has a plurality of identical repeaters which add substantially no delay. Each repeater has at least a first port and a second port connected to a medium, and a third port connected to a slave processor or a master processor. The master processor controls all communications on the medium. A receive multiplexer always applies data on the medium to the processor in the event the data is destined for the processor. A first transmit multiplexer has inputs connected to the second port and the processor, and an output connected to the first port. A second transmit multiplexer has inputs connected to the first port and the processor, and an output connected to the second port. The first and second transmit multiplexers act as a bridge between the first and second ports to pass through data without any variable latency since the data does not have to be buffered.

Abstract: A digital signal detector detects digital signals by only sensing the rising and falling edges of a received digital signal and latches the logic state between the detected edges. Such edges contain very high frequencies that are much higher than the fundamental frequency of the digital signal train. A small high pass filter filters out at least the DC component and the fundamental frequency of the received digital signal. A filtered edge appears as a spike that goes either positive or negative depending on whether the edge is a rising or falling edge. A memory element, such as comprising an RS flip flop, is triggered by the positive and negative spikes. A positive spike triggers the flip flop to output a logical one, and a negative spike triggers the latch to output a logical zero. In this way, the digital signal is recreated without the original digital signal itself being required to pass through the high pass filter.

Abstract: A driver for driving a plurality of light emitting diodes (LEDs) is formed of a plurality of LED controllers connected in series between a power supply and a reference voltage. Each controller drives one or more LEDs directly connected to it. Each controller has a voltage input terminal coupled to an output terminal of an adjacent upstream controller, and an output terminal coupled to the voltage input terminal of an adjacent downstream controller. Each controller has a normally-on bypass switch coupled between its voltage input terminal and the voltage input terminal of the adjacent upstream controller. The bypass switch completely bypasses the adjacent upstream controller when the adjacent downstream controller detects that its input voltage is below a threshold insufficient to drive the LED in the adjacent upstream controller. The bypass switch is turned off if the voltage is above the threshold.

Abstract: A method in a network device implements source address filtering, including gateway address filtering, to enable network devices to be configured in a true Ethernet ring network. By implementing source address filtering or source address filtering with gateway address filtering, a true ring network can be formed using Ethernet protocols where all the links between the network devices in the ring are active paths while avoiding data packets being switched endlessly around the ring. In one embodiment, a data packet in the true ring network is terminated when the source address of the data packet matches the local address of the network device. In another embodiment, a data packet in the true ring network is terminated when the source address of the data packet matches the address of the gateway switch connected to the network device.

Abstract: A digital signal detector detects digital signals by only sensing the rising and falling edges of a received digital signal and latches the logic state between the detected edges. Such edges contain very high frequencies that are much higher than the fundamental frequency of the digital signal train. A small high pass filter filters out at least the DC component and the fundamental frequency of the received digital signal. A filtered edge appears as a spike that goes either positive or negative depending on whether the edge is a rising or falling edge. A memory element, such as comprising an RS flip flop, is triggered by the positive and negative spikes. A positive spike triggers the flip flop to output a logical one, and a negative spike triggers the latch to output a logical zero. In this way, the digital signal is recreated without the original digital signal itself being required to pass through the high pass filter.

Abstract: A Schottky or PN diode is formed where a first cathode portion is an N epitaxial layer that is relatively lightly doped. An N+ buried layer is formed beneath the cathode for conducting the cathode current to a cathode contact. A more highly doped N-well is formed, as a second cathode portion, in the epitaxial layer so that the complete cathode comprises the N-well surrounded by the more lightly doped first cathode portion. An anode covers the upper areas of the first and second cathode portions so both portions conduct current when the diode is forward biased. When the diode is reverse biased, the depletion region in the central N-well will be relatively shallow but substantially planar so will have a relatively high breakdown voltage. The weak link for breakdown voltage will be the curved edge of the deeper depletion region in the lightly doped first cathode portion under the outer edges of the anode. Therefore, the N-well lowers the on-resistance without lowering the breakdown voltage.

Abstract: A method in an Ethernet controller for allocating memory space in a buffer memory between a transmit queue (TXQ) and a receive queue (RXQ) includes allocating initial memory space in the buffer memory to the RXQ and the TXQ; defining a RXQ high watermark and a RXQ low watermark; receiving an ingress data frame; determining if a memory usage in the RXQ exceeds the RXQ high watermark; if the RXQ high watermark is not exceeded, storing the ingress data frame in the RXQ; if the RXQ high watermark is exceeded, determining if there are unused memory space in the TXQ; if there are no unused memory space in the TXQ, transmitting a pause frame to halt further ingress data frame; if there are unused memory space in the TXQ, allocating unused memory space in the TXQ to the RXQ; and storing the ingress data frame in the RXQ.

Abstract: A current sense device for a power transistor is described. The power transistor is formed in a cellular structure including a cellular array of transistor cells. The current sense device includes multiple transistor cells in the cellular array of transistor cells of the power transistor being used as sense transistor cells. The sense transistor cells are evenly distributed throughout the cellular array where the source terminal of each sense transistor cell is electrically connected to a first node through a metal line in the first metal layer and through a metal line in the second metal layer where the metal lines are electrically isolated from the metal lines connecting the transistor cells of the power transistor. The sense transistor cells measure a small portion of the current flowing through the power transistor based on the size ratio of the current sense device and the power transistor.

Abstract: A method for providing adaptive compensation for an electrical circuit where the electrical circuit includes an inductor-capacitor network connected in a feedback loop being compensated by a first compensation capacitance value and a second compensation capacitance value defining the frequency locations of two compensation zeros includes: measuring the inductance value of the inductor; when the inductance value is greater than a first threshold value, increasing the first and second compensation capacitance values so that the frequency locations of the two compensation zeros are adjusted for compensating the poles introduced by the first inductor and the first capacitor; and when the inductance value is less than the first threshold value, decreasing the first and second compensation capacitance values so that the frequency locations of the two compensation zeros are adjusted for compensating the poles introduced by the first inductor and the first capacitor.

Abstract: A method is described for converting an existing die, originally designed for a non-chip-scale package, to a chip-scale package die, where the die's bonding pads are located in positions within a defined grid of candidate positions. In the first step, the die's layout, comprising its outer boundaries and areas needed to be electrically connected to bonding pads, are shifted relative to a grid of candidate positions for the bonding pads until an optimal alignment is identified. Bonding pads positions on the die are then selected corresponding to optimum grid positions within the outer boundaries of the die. The die is then fabricated using the original masks to form at least the semiconductor regions and using a new set of masks for defining the new locations of the bonding pads for the chip-scale package. The chip-scale package is then bonded to a PCB using chip-scale package technology.