Patent Application
20200012302
The disclosure relates to a low-dropout regulator having regulated output voltage spikes, particularly when an output current of the low-dropout regulator is increased.
A low-dropout regulator (LDO) is a DC linear voltage regulator that can regulate the output voltage even when the supply voltage is very close to the output voltage. The LDO provides a regulated output voltage at an output node that may be used to supply a load. An LDO usually comprises an output current branch arranged between a supply potential provided from a supply line and an output node of the LDO to provide the regulated output voltage. The supply line is coupled to a supply source to provide the supply potential at the supply line.
In some applications, it is required that the LDO does not provide a very large change in the output current that the LDO takes from the supply source to be delivered to the load. In an application where the LDO is supplied by means of a long cable, or when a large coil is present on the supply line, it is very important to minimize the supply current derivative. Especially, in the presence of a very small supply cap, the derivate of the output current is responsible for large voltage spikes at the coil terminal of the supply cap.
The output current branch comprises an output driver to provide an output current at the output node, when a load is connected to the output node. The output driver may be configured as a power transistor having a control connection, for example a gate connection, to apply a control voltage for controlling the conductivity of the power transistor. In order to minimize a derivative of the supply current, the control connection/gate connection of the power transistor may be charged/discharged under a slew rate limitation of the control voltage. Hence, the current supply derivative is limited and, in the case of a large coil on the supply line, the supply line is less disturbed.
The slew rate limitation of the control voltage by which the control connection of the output driver is charged/discharged is especially reasonable in the case of large output currents. On the other hand, in the case of a small output current the derivative of the supply current is very small and no significant disturbance on the supply line is observed. Contemporarily, the output driver is not very sensitive at gate regulation for light load currents. As a consequence, large spikes affect the regulated voltage at the output node after a huge transient of the output current towards higher values.
A faster response of the output driver would reduce the spikes at the regulated output voltage but emphasize the supply current variations. Optimization is not possible as the supply stress depends on the power device biasing point.
This is because the power device transconductance is bigger at larger output currents.
It is desired to provide a low-dropout regulator having reduced regulated output voltage spikes, when the output current of the LDO changes.
A low-dropout regulator having reduced regulated output voltage spikes, if a change of the output current occurs, is specified in claim 1.
The low-dropout regulator comprises an output node to provide a regulated output voltage and an output current branch being arranged between a supply line to provide a supply potential and the output node. The output current branch comprises an output driver to provide an output current at the output node. The output driver has a control connection to apply a control voltage. The output driver is configured to be operated with a different conductivity in dependence on the control voltage. The low-dropout regulator further comprises an input amplifier stage to provide the control voltage to the control connection of the output driver. The input amplifier stage is configured to provide the control voltage with a different slew rate in dependence on an increase or decrease of the output current.
In the presence of an output current/load current transient at the output node of the LDO, the largest current spikes at the supply line are generated when the output/load current tends to decrease instead of increasing, independently from the implementation of the LDO. On the contrary, the spikes of the regulated output voltage at the output node are fairly dependent on the LDO architecture but are generally much larger when the output/load current increases its value. The reason for this is the transconductance of the output driver, for example the power transistor, which increases with the output/load current.
In this way, given the same ramp at the control connection of the output driver, for example a gate connection of a transistor, the achieved current variation at the supply line is larger when the output/load current is bigger, while it becomes nearly negligible when the output/load current is in the lowest range and the response of the LDO, for example the transistor arranged in the output current path, is too slow with consequently large spikes at the regulated output voltage.
The presented LDO is configured to increase the slew rate of the control voltage, for example the slew rate of a gate voltage ramp applied to a gate terminal of the transistor of the output driver, when the output/load current is small. In terms of supply-induced disturbances this is not detrimental because the associated supply current derivative remains small enough, but it helps remarkably to reduce the spikes at the regulated voltage at the output node of the LDO, as this is the right condition for them to occur.
According to a first embodiment of the input amplifier stage, the input amplifier stage 30 may comprise a single amplifier circuit 100 having an output side O100 that is directly connected to the control connection G20 of the output driver 20. In this case the amplifier circuit 100 controls the application of the control signal Vc for changing the conductivity of the output driver 20. The input amplifier stage 30 and, in particular, the input amplifier circuit 100 is supplied by the supply potential VDD that is delivered by the supply line Vsupply. The input amplifier circuit 100 has an input side I100 to apply a differential input signal Vin. The input amplifier circuit 100 has an input connection E100a to apply a reference signal Vref and an input connection E100b to apply a feedback signal Vfb. The input signal Vfb is derived from the regulated output voltage Vreg by a feedback net comprising a voltage divider. The voltage divider comprises the resistors 80 and 90.
According to a second embodiment of the input amplifier stage shown in
The buffer circuit 200 has an input side I200 that is connected to the output side O100 of the input amplifier circuit 100. The input amplifier circuit 100 provides the output signal OS that is applied to the input side I200 of the buffer circuit 200. The input amplifier stage 30 is configured such that the buffer circuit 200 controls the application of the control signal Vc to control the output driver 20 by generating a control current Ic at an output side O200. The output side O200 of the buffer circuit 200 is connected to the control connection G20 of the output driver 20. As shown in
The input amplifier stage 30 is configured such that the control connection G20 of the output driver 20, for example the gate connection of the power transistor, is charged/discharged under a slew rate limitation. The input amplifier circuit 100 and/or the buffer circuit 200 provides the charge/discharge control current Ic such that the slew rate of the control voltage Vc at the control connection G20 of the output driver 20 is limited. That means that the input amplifier circuit 100 and/or buffer circuit 200 prevents the control voltage Vc, for example a gate-source voltage of the transistor 20, from increasing too fast so that the output current Iout also cannot vary too fast. As a consequence, a moderately safe control over the supply current variation is achieved. Hence, the current supply derivative is limited and, in case of a large coil connected to the supply potential VDD, the supply line Vsupply is less disturbed.
Despite being less precise, being dependent on temperature, operating conditions and process, the main advantage of the open loop approach of the LDO shown in
As explained above, the buffer circuit 200 can be eliminated so that the input amplifier stage 30 only comprises the input amplifier circuit 100 that directly drives the capacitor 70 and the control connection G20 of the output driver 20 with similar slew rate limitations of the control voltage Vc. The advantage offered by splitting the input amplifier stage 30 so that the input amplifier stage 30 comprises the input amplifier circuit 100 and the buffer circuit 200 is to design the transconductance of the input amplifier circuit 100 independently versus any slew rate concern to ensure better noise and offset performances. Moreover, thanks to the large gain from the input amplifier circuit 100, the buffer circuit 200 undergoes the desired slew rate limitations more easily, even in the presence of small spikes of the regulated output voltage Vreg at the output node O.
Due to the slew rate limitation of the control voltage Vc, the embodiment of the LDO shown in
According to the embodiment shown in
In order to prevent large voltage spikes of the output voltage Vreg in the case of a small output/load current Iout, the input amplifier stage 30 shown in
According to the embodiment of the LDO shown in
Assuming the output driver 20 being configured as an N-MOS transistor, as shown in
In order to realize that the increase of the output/load current Iout takes place with a larger slew rate of the control voltage Vc in comparison to the load/output current decrease, the current mirror circuit 210 of the buffer circuit 200 is configured having a gain K superior to 1. This is because a lower load/output current Iout is less critical than a high one in terms of supply-induced disturbance and faster variations of the control voltage Vc are better tolerated. Providing the current mirror circuit 210 with a gain K superior to 1 makes the pull-up current equal to K*I_tail and keeps the pull-down contribute at I_tail.
The consequent offset from the buffer circuit 200 is negligible, being divided by the gain of the input amplifier circuit 100, if referred to the LDO input. If the input amplifier stage 30 only comprises the input amplifier circuit 100 and the buffer circuit 200 is skipped, the offset is eliminated by mismatching the input differential pair by the same K ratio. According to the embodiment of the LDO 1 shown in
The embodiment of the LDO as shown in
That means that in terms of supply-induced disturbances caused by a variation of the control voltage Vc at the control connection G20, the associated supply control derivative remains small enough and is still acceptable when the output current Iout is large. On the other hand, when the output current Iout is low, the LDO shows a fast response caused by the increased slew rate of the control voltage Vc so that spikes of the regulated output voltage are reduced.
The LDO 2 comprises the output current branch 10 arranged between the supply line Vsupply to provide the supply potential VDD and the output node O to provide the regulated output voltage Vreg. The output current branch 10 comprises the output driver 20 to provide the output current Iout at the output node O. The LDO further comprises the input amplifier stage 30 to provide the control voltage Vc at the control connection G20 of the output driver 20 to control the conductivity of the output driver 20. The input amplifier stage 30 comprises the amplifier circuit 100 and the buffer circuit 200. The output side O100 of the input amplifier circuit 100 is connected to the input side I200 of the buffer circuit 200. The output side O200 of the buffer circuit 200 is connected to the control connection G20 of the output driver 20. The buffer circuit 200 comprises the current mirror circuit 210, the differential input amplifier stage 220 and the bias current source 230. The buffer circuit 200 generates the control voltage Vc at the output side O200. The capacitor 70 is connected to the output side O200 of the buffer circuit 200/the control connection G20 of the output driver 20 and a reference potential VSS.
The input amplifier circuit 100 has an input connection E100a to apply the reference signal Vref and an input connection E100b to apply the feedback signal Vfb being derived from the regulated output voltage Vreg. The buffer circuit 200 receives the output signal OS of the input amplifier circuit 100 at an input connection E200a. An input connection E200b of the buffer circuit 200 is connected to the output side O200 of the buffer circuit 200. The feedback signal Vfb applied to the input connection E100b of the input amplifier circuit 100 is derived from the regulated output voltage Vreg by the voltage divider comprising the resistors 80 and 90.
The LDO 2 comprises a control circuit 300 to control the bias current source 230 of the buffer circuit 200 so that the buffer circuit 200 provides the control voltage Vc at the output side O200 with a larger slew rate, when the output current Iout increases from a first level to a second level. Furthermore, the control circuit 300 controls the bias current source 230 of the buffer circuit 200 so that the buffer circuit 200 provides the control voltage Vc at the output side O200 of the buffer circuit 200 with a smaller slew rate, when the output current Iout increases from the second level to a third level. The first level of the output current is smaller than the second level of the output current, and the second level is smaller than the third level. In order to realize the described operation of the LDO, the LDO 2 comprises a current path 40 and a current mirror stage 50.
The current path 40 is connected between the supply line Vsupply to provide the supply potential VDD and the reference potential VSS. The current path 40 comprises a current driver 41 to provide a replica of the output current Iout of the output current branch 10 in the current path 40. The current driver 41 may be configured as a transistor, for example an N-MOS transistor. The current path 40 further comprises a resistor 42 being connected to the supply line Vsupply and being connected in series to the current driver 41. The current driver 41 is connected to the output node O of the LDO. In particular, the source connection of the current driver 41 is connected to the output node O of the LDO and the control connection/gate connection G41 of the current driver/transistor 41 is connected to the output side O200 of the buffer circuit 200. The current driver 41 is connected with its drain connection to the resistor 42.
The current mirror stage 50 is coupled to the current path 40 and the control circuit 300. The control circuit 300 may be configured as a current mirror stage 60. The current mirror stage 50 is coupled to the current mirror stage 60. The current mirror stage 50 is configured to provide a control current I1 in the current mirror stage 50 to control the bias current I_tail of the bias current source 230 of the buffer circuit 200. The current mirror stage 50 comprises a transistor 51 being arranged between the current mirror stage 60 and a node N1 of the current path 40 located between the current driver 41 and the resistor 42. The current mirror stage 50 further comprises a transistor 52 and a current source 53 being arranged in a current path 54 between the supply line Vsupply and the reference potential VSS. The control connections of the transistors 51 and 52 are directly connected to each other and are additionally connected to a node N3 of the current path 54 between the transistor 52 and the current source 53.
According to the embodiment of the LDO 2, the charge and discharge current Ic are obtained from the shared current root/bias current source 230 that generates the bias current I_tail. The bias current source 230 tracks the output current Iout. The bias current source 230 generates the bias current I_tail with a higher value when the output driver 20 is operated in a low conductive state or nearly in the off-state, and it is minimum when the output driver 20 is crossed by the largest foreseen value of the output current Iout.
In this way, both positive and negative supply current derivatives are reduced when the output driver 20 is biased at the control connection G20 by a large charge/discharge control current Ic, a condition which corresponds to the most critical stress of the supply line Vsupply, while they are kept sufficiently large when the charge/discharge control current Ic is small. This corresponds to the most critical condition for the LDO response speed, while it is not significantly affecting the supply line with disturbances.
The shared current root/the bias current I_tail is obtained by mirroring the output current Iout into a replica in such a way that a larger replica makes a smaller value for the bias current I_tail. The current driver 41 and the resistor 42 of the current path 40 are used together with the current mirror stage 50 and the control circuit 300 to sense the output current Iout and to change the bias current I_tail of the bias current source 230 of the buffer circuit 200, or change a bias current directly in the input amplifier circuit 100, if the buffer circuit 200 is omitted.
According to the embodiment of the LDO 2 shown in
In this way, unlike the implementation 1 of the LDO shown in
Regarding the embodiment 2 of the LDO shown in
The LDO 3 comprises the output current branch 10 with the output driver 20 to provide the output current Iout at the output node O. The LDO 3 further comprises the input amplifier stage 30 comprising the input amplifier circuit 100 and the buffer circuit 200. The buffer circuit 200 comprises the current mirror circuit 210, the differential input amplifier stage 220 and the bias current source 230 to provide the bias current I_tail.
The embodiment of the LDO 3 shown in
When compared to the embodiment of the LDO 2 shown in
The replica of the output current Iout is used to vary the current mirror ratio K of the current mirror circuit 210 to further reduce the rising edge/slew rate of the control voltage Vc of the output driver 20 when the output current Iout is getting large. If the voltage drop across the resistor 42 is negligible, the current mirror gets bigger because of the parallel connection of the transistors 212 and 213. On the contrary, if a large replica current flows across the resistor 42, the voltage drop across the resistor 42 puts off the transistor 213. The transistor 213 tends to mirror less current as soon as the transistor 41 drives more current That means that there is no large rise of the control voltage/gate voltage of the output driver 20, if the output current Iout is large.
The voltage drop across the resistor 42 can optionally be used to reduce the mirror gain for charging the control connection G20 of the output driver 20 in the case of a large output current Iout. Due to the minimum number of nodes/devices involved, the embodiment of the LDO 3 shown in
Despite the solutions shown in
| Number | Date | Country | Kind |
|---|---|---|---|
| 17162558.5 | Mar 2017 | EP | regional |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2018/056067 | 3/12/2018 | WO | 00 |
