Generation of analog voltage using self-biased capacitive feedback stage

1. Technical Field

The present application relates to analog voltage generation, and more particularly to low-power generation of many different desired voltages on a nonvolatile memory integrated circuit.

2. Related Art

Many portable electronic systems are critically limited by battery life. For example, users do not like heavy cell phones, but users also do not like their cell phones to run out of power and shut: down. One way to improve battery life is to increase the energy efficiency of the electronics components in the portable system.

As the constraints of low-power integrated circuits have steadily become tighter, the pressure on voltage management has similarly tightened. Power consumption issues, as well as the normal progress of voltage scaling, push designers toward tighter management of voltages on-chip. Accurate power-valid thresholds, or voltage margins which are dynamically determined in a multi-interface chip, are examples of this.

The availability of power island design techniques also means that reference voltages for the particular needs of a particular power island may be demanded.

Modern nonvolatile memory development is turning into one of the most rapidly advancing areas in the semiconductor industry. Memory cell technology itself continues to advance, even within the general areas However, since the transistor operations are inherently not digital, a variety of reference levels are likely to be needed by memory designs. Thus the capability for generation of secondary analog voltages on chip is important.

The bandgap voltage reference circuit is one of the mainstays of analog electronics, and provides a very reliable on-chip reference. However, this circuit topology, in its many variations, is generally limited to one specific output voltage, and consumes power. It is therefore generally preferable not to use more than one bandgap reference circuit per chip.

The case of a low-power nonvolatile memory module is the convergence of the above requirements. Many different voltages must be made available on chip, and yet static power consumption in the various voltage generation stages is extremely undesirable.

As shown in FIG. 1 , a current source, mirrored from elsewhere on chip, can be combined with a voltage-drop or ohmic element to give a reference voltage output. However, this approach has a constant power loss of the regulated voltage times the reference current squared.

A pending U.S. patent application (U.S. Pat. No. 11/497,465, filed Jul. 31, 2006, titled “Hybrid Charge Pump Regulation”) by the same inventor noted the advantages of a capacitive divider in the feedback loop of a voltage generator circuit. In this example, as illustrated in FIG. 2 , the op amp shown at the right of the diagram drives the charge pump shown at the top left to maintain the output at a level which is equal to the reference voltage Vref times the capacitive ratio.

In the circuit of FIG. 2 , the “div” node is pulled down to ground during the initial phase. After the initial phase the op amp drives the charge pump to make node div equal to Vref, so the output voltage is driven towards Vref times the capacitor ratio (C 1 +C 2 )/C 2 . Note that the output voltage is directly dependent on the capacitor ratio, so process variation, geometric effects, and parasitic capacitance effects can all affect the output voltage.

The present application discloses a significant improvement to that circuit.

The present application discloses new approaches to analog voltage generation using active control without DC current consumption. A conventional leg sets an initial correct voltage value on an output terminal, and a differential stage (e.g. an op amp) is placed in a metastable state at this correct voltage value. The result of this is surprising: the differential stage's connection during startup assures that it will be at its optimal operating point.

The disclosed innovations, in various embodiments, provide one or more of at least the following advantages;

Reduced static current drain More accurate analog voltage generation Reduced power consumption Simpler system architecture Ability to provide a larger number of voltage levels within a given power budget Smaller layout. This is possible because the ratio of the capacitors themselves is not as critical in determining the operating point, according to the present inventions, so it is no longer necessary to use very large capacitors if a precision ratio is required.

The disclosed inventions will be described with reference to the accompanying drawings, which show important sample embodiments of the invention and which are incorporated in the specification hereof by reference, wherein:

FIG. 1 schematically shows a conventional reference voltage generation stage with large static power consumption.

FIG. 2 shows a circuit, having the same inventor as named in the present application, which uses a differential stage to drive an output note to a desired level, which is set by a ratio of capacitors and a reference voltage.

FIG. 3 shows a first sample embodiment of a new circuit, in which a conventional stage is used to set the output voltage, while a differential stage self-tunes itself by connection to the correct output voltage. Note that the output node is not necessarily connected to any external, load at this time.

FIG. 4 shows how two complementary drivers can be used to provide voltage control with, a deadband.

FIG. 5 shows how an op amp has multiple possible operating curves, depending on the DC offset of that particular op amp; and

FIG. 6 shows how the self tuning effect of a circuit like that of FIG. 3 sets a desired operating point, and eliminates the bad effects of DC offset and capacitor ratio.

The numerous innovative teachings of the present application will, be described with particular reference to presently preferred embodiments (by way of example, and not of limitation).

FIG. 3 shows a first: sample embodiment of a new circuit, in which a conventional, stage is used to set the output voltage, while a differential stage self-tunes itself by connection to the correct output voltage. Note that the output node is not necessarily connected to any external load at this time.

During the startup phase (SW driven high), the reference current Iref and the load R drive the output node to approximately IrefR. (If signal SW is well above this value, there will not be a diode drop.) Because the output of the op amp is now connected back to its input (and to node div), the op amp will drive Vdiv to be equal to Vref. Note that SW also gates a PMOS, to assure that the op amp is not able to drive the output node at this time.

After the startup phase (SW low), the IrefR stage is shut off and disconnected, so its power consumption stops. The op amp now drives a pull-up transistor to keep Vdiv=Vref, and hence can respond to changing loads on the output line very rapidly. Since Vdiv was equal to Vref during initialization, and since variations on the output node should be cancelled out by the op amp and current source, the output node will stay at: the same value it had during initialization. Thus the output voltage is no longer dependent on the ratio of capacitors.

The circuit of FIG. 3 shows a pull-up configuration, so this circuit would not be able to offset overvoltage at the output. One way of limiting voltage to a desired range, as shown in FIG. 4 , is to combine a pull-up active stage with an analogous pull-down active stage. However, it is preferable that the two stages do not have the same setpoint, for stability.

FIG. 5 shows a family of op amp transfer curves. The ideal op amp can be thought of as an infinite-gain differential amplifier, so the ideal transfer curve would show a step function at the output=input point. However, a realistic op amp transfer curve will have some curvature, as do the three parallel curves shown.

These curves are parallel because they illustrate an inherent difficulty of op amps, namely uncertain DC offset. As the three oval blobs show, merely setting the input voltage does not tell us which operating curve is the correct one. This makes a difference in the circuit of FIG. 2 , because the DC offset voltage, multiplied by the capacitor ratio, translates to a DC error component on the output node.

FIG. 6 shows how the self tuning effect of a circuit like that of FIG. 3 sets a desired operating point, and eliminates the bad effects of DC offset and capacitor ratio. The dotted line shows how the op amp has been forced to a metastable starting point on its operating curve; this assures that the op amp will respond optimally to transient loading on the output node. The uncertain DC offset of the op amp has been cancelled out.

Modifications and Variations

As will be recognized by those skilled in the art, the innovative concepts described in the present application can be modified and varied over a tremendous range of applications, and accordingly the scope of patented subject matter is not limited by any of the specific exemplary teachings given. It is intended to embrace all such alternatives, modifications and variations that fall within the spirit and broad scope of the appended claims.

For example, additional feedback relations can be added into the illustrated circuit topology if desired, or other conditioning superimposed if desired.

For another example, pass gates can be used if needed to avoid diode drops.

For another example, other circuit topologies can be used to set the correct output value during the precharge phase.

For another example, the timing relations do not have to be as simple as those described above. Optionally some skew and/or gap in timing between the two preset gates and start of steady-state driver operation, or the different transistors which are all shown gated by SW can be skewed from each other.

For another example, the precharge phase can be run repeatedly, to avoid long-term drift in the output voltage,

None of the description in the present application should be read as implying that any particular element, step, or function is an essential element which must be included in the claim scope: The scope of the patented subject matter is defined only by the allowed claims. Moreover, none of these claims are intended to invoke paragraph six of 35 USC section 112 unless the exact words “means for” are followed by a participle.

The claims as filed are intended to be as comprehensive as possible, and no subject matter is intentionally relinquished, dedicated, or abandoned.