Voltage regulators can generally only compensate for rapid transient load changes to a limited extent, i.e. the output voltage thereof experiences deviations from the control voltage, so-called transient deviations, depending on the extent of the load change. The implemented control loops compensate for these control deviations generally over time. In the case of switching converters, this is particularly problematic since they are usually operated in various operating modes for optimizing their efficiency. Thus, a distinction is generally drawn between the discontinuous mode (DCM) and the continuous mode (CCM). In this case, the transition between the modes can take place continuously or in a manner triggered by thresholds by virtue of measuring a value proportional to the load current. The mentioned control deviations have a particularly negative effect in the case of switching converters since, for example on transition from a low load to a high load, the speed at which the connected coil can follow the load current changes is limited by its inductance and the speed of the implemented control loop. It should be noted that the implemented control loops generally need to meet stability criteria which make these control loops slow and therefore the transient deviation unnecessarily great.
The principle illustrated here makes it possible to compensate for transient deviations in accordance with the physical limitation of the coils used outside the conventional control loops. The physical and electrical properties of the components used in the system are used to detect and compensate for load current changes.
The principle illustrated is not restricted to an arrangement for generating one voltage. Such arrangements can also generate several voltages, for example two, three or four.
For optimum system performance, provision is made for an automatic reproduction of the voltage amplitude or phase angles to take place without the need for a communication with a user.
The principle illustrated enables improved compensation of transient deviations of the control voltage (Vout). For this purpose, a control signal is provided which engages in the conventional control loop even before said control loop can correct the control deviations. Provision is made for this control signal to end the monitoring of the control voltage at a favorable point in time, and this monitoring is then resumed again by the conventional control loop.
In the associated illustrations:
The principle is illustrated using the example of a 4-phase digital DC/DC controller. It is not restricted to this embodiment. The illustrated method can also be used for single-phase digital controllers or for analog multi-phase and single-phase controllers.
The mode of operation is illustrated with reference to a block circuit diagram in
It is assumed that at the beginning the DC-to-DC converter operates in a DCM/PFM operating mode, wherein a low load current flows at the output and/or is operated with fewer than the maximum number of possible phases.
For the case where the load current Iload increases quickly, the output voltage Vout will become correspondingly lower. This voltage dip is illustrated at time “t1” in
If the voltage Vout, which is applied to the comparator shown in
By virtue of this operation, the drop in the voltage Vout is compensated for and the voltage increases again, as is illustrated in the lower part of
This method becomes more effective the more noticeable the parasitic physically provided resistances (ESR) and inductances (ESL) of the capacitances and the printed circuit board at the output. This is particularly the case in the case of rapid and high load current changes and when using low capacitances.
A second extended method for phase correction will be described below with reference to
As previously, it is assumed that, at the beginning, the DC-to-DC converter operates in a DCM/PFM operating mode, wherein a low load current is connected or operation is with fewer than the maximum number of possible phases.
When the load current Iload increases rapidly, the output voltage Vout becomes correspondingly lower. This voltage dip is shown at time t1 in
A ratio between the voltage change of Vout and its associated change over time in accordance with dVout/dt is formed by means of the disclosed method with a means suitable for this purpose.
The ratio dVout/dt is low in value, as is illustrated by the rise of the function in
By virtue of the generation of the ratio dVout/dt, the compensation of the voltage dip of Vout already begins at time t1, as is illustrated in
In one configuration of both variants, provision is made for parasitic components of the load (ESL and ESR) to be used for generating the control signal, as is illustrated in
For this purpose, a tap at the load is used, as is illustrated in
This application claims priority of provisional application Ser. No. 61/668,691 filed on Jul. 6, 2012, the entire contents of which is hereby incorporated by reference herein.
Number | Date | Country | |
---|---|---|---|
61668691 | Jul 2012 | US |