WO2009155540A1

WO2009155540A1 - Charge pumps with controlled ramp rate

Info

Publication number: WO2009155540A1
Application number: PCT/US2009/048001
Authority: WO
Inventors: Jeff Kotowski; Shane Hollmer; Curtis B. Robinson, Jr.
Original assignee: Monolithic Power Systems, Inc.
Priority date: 2008-06-20
Filing date: 2009-06-19
Publication date: 2009-12-23

Abstract

A charge pump with a plurality of modes is disclosed. The charge pump reduces EMI and ramp rate by operating at an optimal pumping mode. When a non-optimal mode indication signal is monitored, a charge pump controller will change the mode of pumping to the optimal one.

Description

CHARGE PUMPS WITH CONTROLLED RAMP RATE

TECHNICAL FIELD

[0001] The present invention relates to charge-pumps, especially to charge- pumps with controlled ramp rate.

BACKGROUND

[0002] Two issues that are common for any charge-pump based power device are in-rush current at start-up and RF noise.

[0003] When a charge-pump is first enabled, the pumping capacitors and filter capacitors can be fully discharged and the current drawn from these devices may be many more times that of full power steady state. If left unchecked, large drops on the battery supply will occur, which leads to system unreliability. To mitigate this issue most modern charge-pumps have some sort of soft start circuit which limits the current during this period. Even in steady-state, gain changes, load changes, etc. should be accounted for. [0004] Charge-pumps deliver charge to the output load in packets. The rate at which these packets are introduced to the output (di/dt) modulate the EMI generated. A solution is needed to minimize EMI, but do not give away too much pump efficiency.

[0005] Fig. 1 is a common approach placing a control FET in series with the charge-pump. The control FET may be placed at the input or the output. With this device the charge from the battery can be controlled in various ways some using feedback, some not. For example, a), sense input/output current and use sensed information to control gate voltage of FET; b). during startup, ramp gate voltage of FET so that battery current is increased in controlled fashion. [0006] However, there are some major drawbacks of this method. (1 ) The FET is directly in the power path and thus degrading efficiency and performance. (2) The FET must be made large and thus introducing a cost penalty. [0007] Thus there is unmet need to provide a charge pump to overcome the above problems.

BRIEF DESCRIPTION OF THE FIGURES

[0008] Fig.1 illustrates a common approach placing a control FET in series with the charge-pump in the prior art.

[0009] Fig. 2 illustrates a block diagram of a circuit that reduces the ramp rate of a charge pump in accordance with an embodiment of the present invention.

[0010] Fig. 3 illustrates an embodiment of the pre-driver circuit as in Fig. 2. [0011] Fig. 4 illustrates an embodiment of the ramp control circuit as in Fig. 2. [0012] Fig. 5 illustrates an embodiment of the charge pump used in a white LED application.

[0013] Fig. 6 illustrates a block diagram of an embodiment of the regulation as in Fig. 5.

[0014] The figures depict various embodiments of the present invention for purposes of illustration only. One skilled in the art will readily recognize that.

DETAILED DESCRIPTION OF THE INVENTION

[0015] Fig. 2 illustrates a block diagram 100 reducing the ramp rate of a charge pump in accordance with an embodiment of the present invention. As shown in Fig. 2, block diagram 100 comprises a charge pump circuit providing an output voltage, represented by Vout; a ramp control circuit monitoring the ramp rate of the pump output voltage; a pre-driver circuit receiving the output of the ramp control block, represented as IRC, and providing a signal VPD to the charge pump circuit.

[0016] The strength/resistance of a charge pump is not only related to the pumping capacitor size and frequency, but is also dependent upon the resistance of the switches used to switch the capacitor. The loop shown in block diagram 100 controls the pump by controlling how fast the pump switches turn-on which essentially modulates their resistance. The slower the switches turn-on, the higher their effective resistance are, which results in a weaker pump. The ramp control circuit monitors the output ramp-rate and converts this ramp-rate into a current signal. This current is then used in the pre-driver (switch-driver) circuit to control the turn-on of switches and regulate the strength of the pump. If Vout is moving too fast, the pump strength is reduced to limit the average time the driver spends in the ON state, and vice-versa. In addition to limiting the effective time in the ON state, ramping the gate drive gradually turns on the switches and limits the in-rush current.

[0017] Fig. 3 illustrates an embodiment of the pre-driver circuit as in Fig. 2. In one embodiment, it is used for driving a P-channel device, thus the transition from high to low is needed to control. However, one skilled in the art should realize that the pre-driver circuit could be used for driving an N-channel device or other switching devices. As shown in Fig. 3, a transistor MN5 is a current mirror referenced to the current IRC coming from the ramp control circuit. Transistor MN4 is just a select device and MP5 is responsible for low to high transition. The signal VPD is the output of this pre-driver circuit provided to the charge pump circuit.

[0018] Fig. 4 illustrates an embodiment of the ramp control circuit as in Fig. 2. The ramp control circuit generates a current that is proportional to the output ramp- rate. The current mirrors on the left hand side of the schematic set up various bias currents from one known reference current which flows into MN14. Mirror MP34 sets the max slew current for the N-channel switches. They are not used to limit pump strength, but the max value is set based on EMI considerations. MP8 sets the ramp- rate for pumping down where MN41 sets for pumping up. Finally, MN24 sets the initial condition.

[0019] Changes to the output voltage are coupled to the 'rcnode' via the capacitor divider comprised of C14 and C15. The amount of change is weighted against the control current and sensed by devices MN25 or MN26. These sense devices generate a current somewhere between a fraction of the initial condition current and a max slew current (set by MN33). The generated current is multiplied, mirrored and sent to the pre-driver via device MN30. In reference to a negative charge-pump, when pumping down (softstart), starting with a minimum current through MN30 and working up based on slew rate. In this case using MN25 as the sense device. When pumping up (softstop), starting with maximum current through MN30 and working down. In this case using MN26 as the sense device.

[0020] For decrement mode changes, increase the resistance of switch until either distress is detected or the maximum voltage output capability of the reduced mode is reached. If distress is detected first, just stay in the current mode and allow the pump to go back up. If the max voltage is reached first, switch over to the reduced mode without soft-start. The above ramp control circuit can be used to increase switch resistance by continually lowering the control current.

[0021] To control RF noise generated by the charge pump, the same current source control is utilized. The turn-on slew rate of the switches is controlled via current source and a max current is set in the ramp control circuit. By turning the switches on with a controlled slew rate, the di/dt at the output is minimized.

[0022] Controlling the ramp rate of a charge pump pertains to a variable charge pump in a white LED (WLED) driver application and how to automatically control the gain mode. In one embodiment, the gains available are 1 , 3/2, and 2. Note, however, the gains can be extended to any number of gain modes.

[0023] Fig. 5 illustrate an embodiment of charge pumps used in WLED field. The anode of a WLED is connected to an input Vin, while the cathode is connected to the output of the charge pump via a regulation device. In one embodiment, the regulation device is a current source. Vds represents the drop-out voltage of the regulation device. Any extra voltage in the loop will be dropped across this device. IR represents the drop-out of the charge-pump which will be constant for a given load and Vo is the output voltage of the pump. For this application, equations could be built as: in 1 X mode: Vo(ideal)=0V

In 1.5X mode: Vo(ideal)=-0.5Vin In 2X mode: Vo(ideal)=-1Vin [0024] Therefore:

In 1 X mode: Vf=Vin-Vds-IR In 1.5X mode: Vf= 1.5Vin-Vds-IR In 2X mode: Vf=2Vin-Vds-IR

[0025] Assuming IR is for all modes and Vds(min)=Vmin, the above signals should meet the relationship: Vds>0.5Vin+Vmin. When Vds<0.5Vin+Vmin happens, the charge pump mode needs decrement, such as goes to 1X mode from 1.5X mode, or goes to 1.5X from 2X mode.

[0026] As a result, which mode the charge-pump needs to be in to meet the Vf requirement can be determined.

[0027] For increasing the gain mode, just monitoring the regulation device. When current goes out of regulation, increase the gain.

[0028] For regulation to occur in any mode, the charge-pump output minus IR minus Vmin must be greater than Vf. It can be seen that the output voltage from the pump decreases by 0.5Vin for each decrement in gain. This depends on the gain modes available but similar relationships can be developed for any number of modes. Therefore, if the voltage across the current regulation device (Vds) is greater than 0.5Vin, i.e., Vds>0.5Vin, decrease the gain and maintain regulation. Since there is some minimum drop-out associated with the regulation, the decrement can not be at exactly 0.5Vin, but instead at O.δVin plus Vdsat of regulation device at load. This minimum drop-out voltage can be determined through simulation. [0029] Fig. 6 illustrates a block diagram of an embodiment of the regulation as in Fig. 5. However, one skilled in the art should realize that there are multiple alternatives. The counter and mux cycles through the drain voltages, one for each LED being driven. A comparator UO is used to receive a signal (2/5Vs+1/5Vin+2/5Vmin) at its non-inverting input termianl, and receive a signal 2/5Vd at its inverting input terminal. When signal (2/5Vs+1/5Vin+2/5Vmin) is higher than signal 2/5Vd, that is 1/2Vin+Vmin>Vd-Vs, i.e., Vds<0.5Vin+Vmin, the output of the comparator UO is high. A FF and latch block allows each drain voltage to be compared in sequence and ensure that only when all meet the decrement requirement is a DEC signal asserted. Then the gain mode of the charge pump is decreased. [0030] The foregoing description of the embodiments of the invention has been presented for the purpose of illustration; it is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Persons skilled in the relevant art can appreciate that many modifications and variations are possible in light of the above teachings. It is therefore intended that the scope of the invention be limited not by this detailed description, but rather by the claims appended hereto.

Claims

CLAIMSWhat is claimed is:

1. A method of controlling an input current to and electromagnetic interference (EMI) from a charge pump circuit, comprising: providing an output through a charge pump at a certain mode; measuring the ramp rate of the output; whenever the ramp rate of the output indicates the charge pump is at a non- optimal mode, then changing the charge pump to an optimal mode.

2. The method of claim 1 , further comprising converting the ramp rate of the output into a current signal.

3. The method of claim 1 , wherein the mode change is mode decrement.

4. The method of claim 1 , wherein the ramp rate is charging rate.

5. The method of claim 1 , wherein the ramp rate is discharging rate.

6. An apparatus, comprising: a charge pump circuit adapted to regulate a voltage by charge pumping at a mode that satisfies the voltage to provide a regulated voltage, the charge pump circuit having a plurality of modes of pumping; and a charge pump controller, electrically coupled to the charge pump circuit, operable to monitor the ramp rate of the regulated voltage, change the mode of pumping when the ramp rate of the regulated voltage indicates that the charge pump is at a non-optimal mode.

7. The apparatus of claim 6, wherein the charge pump controller converts the ramp rate of the regulated voltage into a current signal after monitoring.

8. The apparatus of claim 6, wherein the charge pump controller comprises a ramp control circuit and a pre-dhver circuit.

9. The apparatus of claim 8, wherein the ramp control circuit monitors the ramp rate of the regulated voltage and converts to a current signal.

10. The apparatus of claim 9, wherein the pre-driver circuit receives the current signal and provides a signal to modulate the resistance of said charge pump circuit.

11. The apparatus of claim 10, wherein the higher the resistance, the weaker the pump gain.

12. The apparatus of claim 6, wherein the charge pump controller comprises a comparator, electrically coupled to compare two signals and provide an indication signal that whether the charge pump is at a non-optimal mode; a counter, electrically coupled to ensure that each drain signal is compared; a FF and latch circuit, electrically coupled to provide a decrement signal.