System and method for automatically tuning a voltage converter

• US 8,670,845 B2
• Filed: 08/31/2010
• Issued: 03/11/2014
• Est. Priority Date: 09/14/2009
• Status: Active Grant

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Dabak, Anand Ganesh, McDonald, Brent
• Primary Examiner(s): WILSON, KIMBERLY LOVEL
• Application Number:
  US12/872,256
• Publication Number:
  US 20110063881A1
• Time in Patent Office: 1,288 Days
• Field of Search: 700/55
• US Class Current: 700/55
• CPC Class Codes: H02M 3/157 with digital control
• CPC Class Codes: H02M 7/53873 with digital control

Abstract

A power converter system is provided, comprising a plant having a plant input and a plant output; and a plant identification filter that receives the plant input and the plant output, and estimates the values of poles and zeros of the plant, wherein the plant identification filter updates the estimates of the poles and zeros, based upon the plant input and the plant output, beginning from an initial state; and a rate at which the plant identification filter updates the estimates of the values of the poles and zeros is slower than a rate at which the plant input and the plant output are received by the plant identification filter.

Claims

1. A power converter system, comprising:

a plant having a plant input and a plant output;

a plant identification filter circuit hardware that receives the plant input and the plant output, and estimates values of poles and zeros of the plant, comprising;

a Kalman filter;

a pair of anti-aliasing, low-pass filters connected in series with the Kalman filter the first of the pair of anti-aliasing, low-pass filters receives the plant input, decimates the plant input by a second integer factor, and outputs the decimated plant input to the Kalman filter, and the second of the pair of anti-aliasing, low-pass filters receives the plant output, decimates the plant output by the second integer factor, and outputs the decimated plant output to the Kalman filter, wherein the plant identification filter circuit hardware updates the estimates of the values of the poles and zeros, based upon the plant input and the plant output, beginning from an initial state, a rate at which the plant identification filter updates the estimates of the values of the poles and zeros is slower than a rate at which the plant input and the plant output are received by the plant identification filter, wherein the Kalman filter updates the estimates of the values of the poles and zeros by determining Kalman coefficients using recursive least squares on the plant input and the plant output, and the rate at which the Kalman filter updates the estimates of the values of the poles and zeros is slower than the rate at which the plant input and the plant output are received by a first integer factor, wherein the first integer factor is set to a first particular integer and the second integer factor is set to second particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is less than a predetermined low zero-pole threshold that is based on the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively, the second integer factor is reset to an integer that is greater than the second particular integer and the first integer factor is reset to an integer that is less than the first particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is greater than a predetermined high zero-pole threshold that is based on the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively, the second integer factor is reset to an integer that is less than the second particular integer and the first integer factor is reset to an integer that is greater than the first particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is neither less than the predetermined low zero-pole threshold nor greater than the predetermined high zero-pole threshold, the second integer factor remains set at the second particular integer and the first integer factor remains set at the first particular integer, and

if a maximum of the updated estimates of the values of the zeros and poles is either less than the low zero-pole threshold or greater than the high zero-pole threshold,

the first of the pair of anti-aliasing, low-pass filters receives the plant input, decimates the plant input by the second integer factor as reset, and outputs the decimated plant input to the Kalman filter, and

the second of the pair of anti-aliasing, low-pass filters receives the plant output, decimates the plant output by the second integer factor as reset, and outputs the decimated plant output to the Kalman filter,

the rate at which the Kalman filter updates the estimates of the values of the poles and zeros is slower than the rate at which the plant input and the plant output are received by the first integer factor as reset.

2. The power converter system according to claim 1, wherein

the predetermined low zero-pole threshold is 10% of the rate at which the plant input and the plant output are received by the plant identification filter, and

the predetermined high zero-pole threshold is 50% of the rate at which the plant input and the plant output are received by the plant identification filter.

3. The power converter system according to claim 2, wherein

the rate at which the plant input and the plant output are received by the plant identification filter is 1 MHz,

the first particular integer is 8 and the second particular integer is 16,

the integer that is greater than the second particular integer is 32 and the integer that is less than the first particular integer is 4,

the integer that is less than the second particular integer is 8 and the integer that is greater than the first particular integer is 16.

4. The power converter system according to claim 1, wherein

the plant is a Buck converter plant, including

one or more passive devices that apply an output voltage to a load,

one or more transistors that, by being turned on or off, regulate an input voltage to the one or more passive devices,

the plant output is the output voltage, and

the plant input is an input voltage that effects whether each of the one or more transistors is turned on or off.

5. A plant identification filter circuit for estimating the values of poles and zeros of a plant, comprising:

a plant reception circuitry that receives, as a plant input, an input to the plant, and further receives, as a plant output, an output from the plant; and

a plant identification circuitry that updates the estimates of the values of the poles and zeros, based upon the plant input and the plant output, beginning from an initial state, the rate at which the plant identification circuitry updates the estimates of the values of the poles and zeros being slower than a rate at which the plant input and the plant output are received by the plant reception circuitry, the plant identification circuitry comprising;

a Kalman filter which updates the estimates of the values of the poles and zeros by determining Kalman coefficients using recursive least squares on the plant input and the plant output and the rate at which the Kalman filter updates the estimates of the values of the poles and zeros is slower than the rate at which the plant input and the plant output are received by a first integer factor;

a pair of anti-aliasing, low-pass filters connected in series with the Kalman filter, the first of the pair of anti-aliasing, low-pass filters receives the plant input, decimates the plant input by a second integer factor, and outputs the decimated plant input to the Kalman filter, and the second of the pair of anti-aliasing, low-pass filters receives the plant output, decimates the plant output by the second integer factor, and outputs the decimated plant output to the Kalman filter, wherein

the first integer factor is set to a first particular integer and the second integer factor is set to second particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is less than a predetermined low zero-pole threshold that is based on the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively, the second integer factor is reset to an integer that is greater than the second particular integer and the first integer factor is reset to an integer that is less than the first particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is greater than a predetermined high zero-pole threshold that is based on the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively, the second integer factor is reset to an integer that is less than the second particular integer and the first integer factor is reset to an integer that is greater than the first particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is neither less than the predetermined low zero-pole threshold nor greater than the predetermined high zero-pole threshold, the second integer factor remains set at the second particular integer and the first integer factor remains set at the first particular integer, and

if a maximum of the updated estimates of the values of the zeros and poles is either less than the low zero-pole threshold or greater than the high zero-pole threshold,

the first of the pair of anti-aliasing, low-pass filters receives the plant input, decimates the plant input by the second integer factor as reset, and outputs the decimated plant input to the Kalman filter, and

the second of the pair of anti-aliasing, low-pass filters receives the plant output, decimates the plant output by the second integer factor as reset, and outputs the decimated plant output to the Kalman filter,

the rate at which the Kalman filter updates the estimates of the values of the poles and zeros is slower than the rate at which the plant input and the plant output are received by the first integer factor as reset.

6. The plant identification filter circuit according to claim 5, wherein

the predetermined low zero-pole threshold is 10% of the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively

the predetermined high zero-pole threshold is 50% of the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively.

7. The plant identification filter circuit according to claim 6, wherein

the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti- aliasing, low-pass filters, respectively, is 1 MHz,

the first particular integer is 8 and the second particular integer is 16,

the integer that is greater than the second particular integer is 32 and the integer that is less than the first particular integer is 4,

the integer that is less than the second particular integer is 8 and the integer that is greater than the first particular integer is 16.

8. A method comprising:

receiving, by a plant identification filter, as a plant input, an input to a plant, and further receiving, as a plant output, an output from the plant;

updating by a Kalman filter in the plant identification filter estimates of values of poles and zeros by determining Kalman coefficients using recursive least squares on the plant input and the plant output, the rate at which the Kalman filter updates the estimates of the values of the poles and zeros is slower than the rate at which the plant input and the plant output are received by a first integer factor;

a pair of anti-aliasing, low-pass filters are connected in series with the Kalman filter, the first of the pair of anti-aliasing, low-pass filters receives the plant input, decimates the plant input by a second integer factor, and outputs the decimated plant input to the Kalman filter, and

the second of the pair of anti-aliasing, low-pass filters receives the plant output, decimates the plant output by the second integer factor, and outputs the decimated plant output to the Kalman filter, wherein

the first integer factor is set to a first particular integer and the second integer factor is set to second particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is less than a predetermined low zero-pole threshold that is based on the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively, the second integer factor is reset to an integer that is greater than the second particular integer and the first integer factor is reset to an integer that is less than the first particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is greater than a predetermined high zero-pole threshold that is based on the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti- aliasing, low-pass filters, respectively, the second integer factor is reset to an integer that is less than the second particular integer and the first integer factor is reset to an integer that is greater than the first particular integer,

if a maximum of the updated estimates of the values of the zeros and poles is neither less than the predetermined low zero-pole threshold nor greater than the predetermined high zero-pole threshold, the second integer factor remains set at the second particular integer and the first integer factor remains set at the first particular integer, and

if a maximum of the updated estimates of the values of the zeros and poles is either less than the low zero-pole threshold or greater than the high zero-pole threshold,

the first of the pair of anti-aliasing, low-pass filters receives the plant input, decimates the plant input by the second integer factor as reset, and outputs the decimated plant input to the Kalman filter, and

the second of the pair of anti-aliasing, low-pass filters receives the plant output, decimates the plant output by the second integer factor as reset, and outputs the decimated plant output to the Kalman filter,

the rate at which the Kalman filter updates the estimates of the values of the poles and zeros is slower than the rate at which the plant input and the plant output are received by the first integer factor as reset.

9. The method according to claim 8, wherein

the predetermined low zero-pole threshold is 10% of the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively, and

the predetermined high zero-pole threshold is 50% of the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively.

10. The method according to claim 9, wherein

the rate at which the plant input and the plant output are received by the first of the pair of anti-aliasing, low-pass filters and the second of the pair of anti-aliasing, low-pass filters, respectively, is 1 MHz,

the first particular integer is 8 and the second particular integer is 16,

the integer that is greater than the second particular integer is 32 and the integer that is less than the first particular integer is 4,

the integer that is less than the second particular integer is 8 and the integer that is greater than the first particular integer is 16.

11. The method according to claim 8, wherein

the plant is a Buck converter plant, including

one or more passive devices that apply an output voltage to a load,

one or more transistors that, by being turned on or off, regulate an input voltage to the one or more passive devices,

the plant output is the output voltage, and

the plant input is an input voltage that effects whether each of the one or more transistors is turned on or off.