Method for calibrating variable delay circuit and a variable delay circuit using the same

• US 6,163,759 A
• Filed: 11/17/1998
• Issued: 12/19/2000
• Est. Priority Date: 11/21/1997
• Status: Expired due to Term

• Current Assignee: Advantest Corporation
• Original Assignee: Advantest Corporation
• Inventors: Kita, Kazumi
• Primary Examiner(s): Hoff, Marc S.
• Assistant Examiner(s): Raymond, Edward
• Application Number:
  US09/193,523
• Time in Patent Office: 763 Days
• Field of Search: 327/261, 327/141, 73/1.01, 73/5, 73/6, 73/1, 73/35, 368/114, 322/18, D10/40, 324/130, 714/724, 702/85, 702/176, 702/1, 702/108, 702/127
• US Class Current: 702/85
• CPC Class Codes: G01R 31/3191 Calibration

Abstract

In a variable delay circuit calibrating method in which the state of connection of M delay stages connected in cascade through multiplexers and weighted differently is controlled by a control signal value to generate a calibrated amount of delay corresponding to a nominal amount of delay D_s which varies in a predetermined minimum nominal delay step d_s, the method comprises the steps of: dividing an amount of delay D_i measured for each given control signal value CC_i by the minimum nominal delay step d_s of a variable delay circuit; calculating first and second errors, R_k =D_i -d_s k and R_k+1 =d_s -R_k, between the value k of an integral part of the resulting quotient and two adjoining nominal amounts of delay D_sk and D_sk+1 ; making a check to determine if the first error R_k is smaller than an error held in a k-th row of a calibration table in correspondence with a nominal set signal value CS=k; if so, writing the first error R_k and the corresponding control signal value CC_i over the existing values in the column of the k-th row of the calibration table; making a check to determine if the second error R_k+1 is smaller than an error held in a (k+1)-th row of the calibration table in correspondence with a nominal setting signal value CS=k+1; and, if so, writing the second error R_k+1 and the corresponding control signal value CC_i over the existing values in the column of the (k+1)-th row of the calibration table. By repeatedly executing these steps for i=0 to i=2^M -1, control signal values, which minimize errors between delay times of the variable delay circuit and the nominal amounts of delay, are generated in O-th to K-th rows of the calibration table in correspondence with the respective nominal setting signal values CS.

Claims

1. A method for calibrating, for a nominal amount of delay D_sk =d_s k, k=0, 1, 2, . . . , K, for each nominal minimum delay step d_s, a variable delay circuit in which the state of a cascade connection of M differently weighted delay stages is selectively controlled by a desired one of control signal values CC_i from O to K to thereby set a desired amount of delay of said variable delay circuit, said M and K being integers equal to or greater than 2, said method comprising the steps of:

(a) predefining in a memory an at least K-row calibration table having columns for holding errors and control signal values which provide the errors;

(b) measuring the amount of delay D_i of said variable delay circuit set by said control signal value CC_i ;

(c) dividing said measured amount of delay D_i by said minimum nominal step d_s and calculating an integral part k of the resulting quotient and first and second errors R_k =D_i -d_sk and R_k+1 =d_s -R_k of said measured amount of delay D_i with respect to two adjoining nominal amounts of delay D_sk and D_sk+1 ;

(d) making a check to determine if said calculated first error R_k is smaller than an error held in a k-th row of said calibration table, and if so, writing said calculated first error R_k and the corresponding control signal value CC_i over previous data in the column of said k-th row;

(e) making a check to determine if said calculated second error R_k+1 is smaller than an error held in a (k+1)-th column of said calibration table, and if so, writing said calculated second error R_k+1 and the corresponding control signal values CC_i over previous data in the column of said (k+1)-th row; and

(f) executing said steps (b) to (e) from i=0 to i=2^M -1, thereby obtaining, in O-th to K-th rows of said calibration table, control signal values which minimize error with respect to said nominal amount of delay.

2. The method of claim 1, wherein said step (s) includes a step of prewriting a fixed value greater than said nominal delay step d_s, as an initial value in the column of error of said calibration table.

3. The method of claim 1 or 2, wherein, if the delay time t_d0 between the input and output of said variable delay circuit, when supplied with a control signal CC₀ for bypassing all of said delay stages, is taken as an offset delay, said step (b) includes a step of measuring the delay time t_d1 between the input and output of said variable delay circuit for each of said control signal values CC_i and obtaining said amount of delay D_i as given by the following equation:

space="preserve" listing-type="equation">D.sub.i =t.sub.d1 -t.sub.d0.

4. The method of claim 1 or 2, wherein an m-th (m=1, 2, . . . , M) one of said delay stages is weighted about 2^m-1.

5. A variable delay circuit which generates a delay calibrated by a control signal value in correspondence with a nominal amount of delay D_sk for each nominal minimum delay step d_s, said variable delay circuit comprising:

M delay stages of differently weighted amounts of delay;

M multiplexers connected to the outputs of said M delay stages, respectively, for selectively outputting inputs to and outputs from said delay stages, said M delay stages and said M multiplexers being connected in cascade;

control signal converting means responsive to a control signal fed thereto, for generating desired control signal values from O to K and applying them to said multiplexers to set in said M delay stages desired amounts of delay corresponding to nominal amounts of delay D_sk =d_s k (k=0, 1, 2, . . . , K) for each nominal minimum delay step, said M and K being integers equal to or greater than 2;

delay time measuring means for measuring the delay time between the input and output of said variable delay circuit upon each change of said control signal values from O to K;

a memory having defined therein an at least K-row calibration table with columns for prestoring an error and a control signal value which provides said error; and

calculating means for;

dividing an amount of delay D_i measured for each given control signal value CC_i by the minimum nominal delay step d_s of a variable delay circuit;

calculating first and second errors, R_k =D_i -d_s k and R_k+1 =d_s -R_k, between the value k of an integral part of the resulting quotient and two adjoining nominal amounts of delay D_sk and D_sk+1 ;

making a check to determine if the first error R_k is smaller than an error held in a k-th row of a calibration table in correspondence with a nominal set signal value CS=k;

if so, writing the first error R_k and the corresponding control signal value CC_i over the existing values in the column of the k-th row of the calibration table;

making a check to determine if the second error R_k+1 is smaller than an error held in a (k+1)-th row of the calibration table in correspondence with a nominal setting signal value CS=k+1;

if so, writing the second error R_k+1 and the corresponding control signal value CC_i over the existing values in the column of the (k+1)-th row of the calibration table; and

repeatedly executing these operations for i=0 to i=2^M -1, whereby control signal values, which minimize errors between delay times of the variable delay circuit and the nominal amounts of delay, are generated in O-th to K-th rows of the calibration table in correspondence with the respective nominal setting signal values CS.

6. The variable delay circuit of claim 5, wherein there is prestored, as an initial value, a fixed value greater than said nominal delay step d_s in the column of error of said calibration table in said memory.

7. The variable delay circuit of claim 5 or 6, wherein , if the delay time t_d0 between the input and output of said variable delay circuit, when supplied with a control signal CC₀ for bypassing all of said delay stages, is taken as an offset delay, said delay time measuring means measures the delay time t_d1 between the input and output of said variable delay circuit for each of said control signal values CC_i and said calculating means obtains said amount of delay D_i as given by the following equation:

space="preserve" listing-type="equation">D.sub.i =t.sub.d1 -t.sub.d0.

8. The variable delay circuit of claim 5 or 6, wherein an m-th (m=1, 2, . . . , M) one of said delay stages is weighted about 2^m-1.