Evaluation circuit and method for detecting and/or locating faulty data words in a data stream Tn

• US 8,060,800 B2
• Filed: 10/22/2004
• Issued: 11/15/2011
• Est. Priority Date: 10/24/2003
• Status: Active Grant

• Current Assignee: Infineon Technologies AG
• Original Assignee: Infineon Technologies AG
• Inventors: Mattes, Heinz, Leininger, Andreas, Sattler, Sebastian, Goessel, Michael
• Primary Examiner(s): Gaffin; Jeffrey A
• Assistant Examiner(s): MCMAHON, DANIEL F
• Application Number:
  US10/577,288
• Publication Number:
  US 20080040638A1
• Time in Patent Office: 2,580 Days
• Field of Search: 714/724, 714/732, 714/733, 714/734
• US Class Current: 714/732
• CPC Class Codes: G01R 31/31703 Comparison aspects, e.g. signature analysis, comparators concerning scan tests G01R31/318566; concerning testers G01R31/3193
• CPC Class Codes: G01R 31/31932 Comparators

Abstract

An evaluation circuit and method for detecting faulty data words in a data stream is disclosed. In one embodiment the evaluation circuit according to the invention includes a first linear automaton circuit and also a second linear automaton circuit connected in parallel, each having a set of states z, which have a common input line for receiving a data stream T_n. The first linear automaton circuit and the second linear automaton circuit are designed such that a first signature and a second signature, respectively, can be calculated. Situated downstream of the two linear automaton circuits are respectively a first logic combination gate and a second logic combination gate, which compare the signature respectively calculated by the linear automaton circuit with a predeterminable good signature and output a comparison value.

Claims

1. An evaluation circuit for detecting and/or locating faulty data words in a data stream T_n comprising:

a first linear automaton circuit and a second linear automaton circuit connected in parallel, each having a set of states, wherein the first linear automaton circuit and the second linear automaton circuit each have inputs that are commonly connected for receiving a data stream T_n comprising n successive data words y(1), . . . , y(n) each having a width of k bits, k>

1,

wherein the first linear automaton circuit can be described by the following equation

z(t+1)=Az(t)⊕

y(t)

wherein the second linear automaton circuit can be described by the following equation

z(t+1)=Bz(t)⊕

y(t)

where z represents state vectors and A and B represent the state matrices of the linear automaton circuits, where the state matrices A and B can be inverted, and where a dimension L of the state vectors z is ≧

k, wherein A≠

B,

the first linear automaton circuit and the second linear automaton circuit are designed such that a first signature and a second signature, respectively, is calculated of each data word of the n successive data words y(1), . . . , y(n),

L first logic combination gates arranged downstream of the first linear automaton circuit and also L second logic combination gates arranged downstream of the second linear automaton circuit,

the logic combination gates are designed such that the signature respectively calculated by the linear automaton circuit can be compared with a predeterminable good signature and a comparison value can be output.

2. The evaluation circuit as claimed in claim 1, comprising wherein the logic combination gates are present as exclusive-OR gates whose first inputs are respectively connected to the outputs of the associated linear automaton circuit (L1, L2) and to whose second inputs good signatures can be applied.

3. The evaluation circuit as claimed in claim 1, comprising wherein arranged upstream of the first linear automaton circuit is a first coder, that codes the data word y(i) having the data word length of k bits into a coded data word u¹(i), u¹(i)=Cod1 having the word width of K1 bits, for i=1, . . . , n, and where Cod1 represents the coding function of the first coder.

4. The evaluation circuit as claimed in claim 3, comprising wherein the following holds true for the coding function of the first coder:

5. The evaluation circuit as claimed in claim 3, comprising wherein arranged upstream of the second linear automaton circuit is a second coder, which codes the data word y(i) having the data word length of k bits into an encoded data word u²(i), u²(i)=Cod2(y(i)) having the word width of K2 bits, for i=1, . . . , n, and where Cod2 represents the coding function of the second coder.

6. The evaluation circuit as claimed in claim 5, comprising wherein the following holds true for the coding function of the second coder:

Cod2(y′

(i))=u²(i)⊕

f₂(e(i)),

or

Cod2(y′

(i))=Cod2(y(i)⊕

e(i))=Cod2(y(i))⊕

f₂(e(i))

where a function f₂<sup>−

1</sup> where

f₂<sup>−

1</sup>(f₂(e))=e

exists for all binary data words e having the word width k which may occur as errors of a data word, where e denotes a faulty data word of the data stream T_n.

7. The evaluation circuit as claimed in claim 5, comprising wherein that the word width K1 of the data words u¹(i) coded by the first coder is equal to the word width K2 of the data words u²(i) coded by the second coder.

8. The evaluation circuit as claimed in claim 5, comprising wherein the first coder matches the second coder with regard to its construction and its function.

9. The evaluation circuit as claimed in claim 5, comprising wherein the word width K1 of the data words u¹(i) coded by the first coder and the word width K2 of the data words u²(i) coded by the second coder are in each case equal to the word width k of the data words y(1), . . . , y(n) of the data stream T_n.

10. The evaluation circuit as claimed in claim 5, comprising wherein the coding functions Cod1 and Cod2 of the first coder and of the second coder are designed as follows:

Cod1(y₁(i),y₂(i), . . . , y_k(i))=P1(y₁(i),y₂(i), . . . , y_k(i), 0, . . . ,

     0)

Cod2(y₁(i),y₂(i), . . . , y_k(i))=P2(y₁(i),y₂(i), 0, . . . ,

     0)

for i, 1, . . . , n

where the number of zeros situated at the end of P1(y₁(i), y₂(i), . . . , y_k(i), 0, . . . ,

     0) is equal to (K1−

k), where the number at the end of P2(y₁(i), y₂(i), . . . , y_k(i), 0, . . . ,

     0) is equal to (K2−

k), and where P1 represents an arbitrary permutation of the K1 components of (y₁(i), y₂(i), . . . , y_k(i), 0, . . . ,

     0) and P2 represents an arbitrary permutation of the K2 components of (y₁(i), y₂(i), y_k(i), 0, . . . ,

     0).

11. The evaluation circuit as claimed in claim 3, comprising wherein the coding functions Cod1 and Cod2 of the first coder and of the second coder are designed as follows:

Cod1(y₁(i),y₂(i), . . . , y_k(i))=P1(y_i(i),y₂(i), . . . , y_k(i),b₁¹ . . . , b_{K1 k}¹)

Cod2(y₁(i),y₂(i), . . . , y_k(i))=P2(y₁(i),y₂(i), . . . , y_k(i),b₁² . . . , b_{K2 k}²)

where b₁¹, . . . , b<sub>K1−

k</sub>¹, b₁², . . . , b<sub>K2−

k</sub>² ε

{0.1}, and where P1 and P2 represent arbitrary permutations.

12. The evaluation circuit as claimed in claim 3, comprising wherein the coding function Cod1 of the first coder is designed such that it realizes a linear block code, f₁=Cod1.

13. The evaluation circuit as claimed in claim 5, comprising wherein the coding function Cod2 of the second coder is designed such that it realizes a linear block code, f₂=Cod2.

14. The evaluation circuit as claimed in claim 1, comprising wherein the state matrix A of the first linear automaton circuit and the state matrix B of the second linear automaton circuit are related to one another as follows:

B=Aⁿ

where n≠

1.

15. The evaluation circuit as claimed in claim 1, comprising wherein the state matrix B of the second linear automaton circuit is equal to the inverted state matrix A⁻

1 of the first linear automaton circuit.

16. The evaluation circuit as claimed in claim 1, comprising wherein the first linear automaton circuit is designed as a linear feedback shift register and the second linear automaton circuit is designed as an inverse linear feedback shift register, both linear automaton circuits having a parallel input.

17. The evaluation circuit as claimed in claim 1, comprising wherein the first linear automaton circuit is designed as a linear feedback, K1-dimensional multi-input shift register or the second linear automaton circuit is designed as a linear feedback, multi-input shift register.

18. The evaluation circuit as claimed in claim 17, comprising wherein the multi-input shift register have a primitive feedback polynomial of maximum length.

20. The evaluation circuit as claimed in claim 1, comprising wherein the evaluation circuit is monolithically integrated on an integrated circuit, each data word having a data word length of k.

21. A loadboard for receiving at least one needle card for testing integrated circuits or having at least one test socket for testing integrated circuits or for connecting a handler to a tester of integrated circuits, the loadboard having an evaluation circuit as claimed in claim 1.

22. A needle card for testing integrated circuits, in which an evaluation circuit as claimed in claim 1 is integrated.

23. A tester for testing integrated circuits having the following features:

the tester is provided with a plurality of instruments for generating signals or data streams and with a plurality of measuring sensors, in particular for currents and voltages,

the tester has a loadboard which is provided for receiving at least one needle card for testing integrated circuits and/or for connecting a handler to a tester of integrated circuits and/or which is equipped with at least one test socket for testing integrated circuits, and

the tester has an evaluation circuit as claimed in claim 1.

19. A method for detecting and/or locating faulty data words in a data stream T_n, the method having the following method steps of:

inputting data words y(1), . . . , y(i−

1), y′

(i), y(i+1), . . . , y(n) of a data stream T_n into a first coder, each data word having a width of k bits, k>

1,

encoding the data words y(1), . . . , y(n) into coded data words u¹(1), . . . , u¹(n) having a word width K1 where K1≧

k by means of the coding function Cod1 of the first coder,

inputting the coded data words u¹(1), . . . , u¹(i−

1), u¹′

(i) or u¹(i), u¹(i), . . . , u¹(n) into the inputs of a first linear automaton circuit, which is described by the automaton equation;

z¹(t+1)=A·

z¹(t)+u¹(t)

where t is an instant in time, z¹ represents a K1-dimensional state vector and A represents a K1×

K1 state matrix, and where the state matrix A can be inverted,

processing the coded data words u¹(1), . . . , u¹(i−

1), u¹′

(i) or u¹(i), u¹(i), . . . , u¹(n) by means of the first linear automaton circuit, the first linear automaton circuit,

undergoing transition to the state z¹(n+1)=S₁(L1, y(1), . . . , y(i−

1), y(i), y(i+1), . . . , y(n)) if no error can be detected in the case of the coded data words u¹(1), . . . , u¹(i−

1), u¹(i), u¹(i+1), . . . , u¹(n),

undergoing transition to the state z¹′

(n+1)=S₁(L1, y(1), . . . , y(i−

1), y′

(i), y(i+1), . . . , y(n)) if an error is present at least in the case of the i-th position of the coded data words u¹(1), . . . , u¹(i−

1), u¹′

(i), . . . , u¹(n),

the signature of an error-free data stream T_n being designated by S(L1, y(1), . . . , y(i−

1), y(i), y(i+1), . . . , y(n)) and the signature of a faulty data stream T_n being designated by S(L1, y(1), . . . , y(i−

1), y′

(i), y(i+1), . . . , y(n)),

checking the determined signature of the data stream T_n and continuing with method step a) for further data streams T_n if the determined signature of the data stream T_n is the signature of an error-free data stream T_n,

inputting the data words y(1), . . . , y(i−

1), y′

(i), . . . , y(n) of the data stream T_n in a second coder,

coding the data words y(1), . . . , y(i−

1), y′

(i), y(i+1), . . . , y(n) to coded data words u²(1), . . . , u²(i−

1), u²′

(i) or u²(i), u²(i), . . . , u²(n) having the word width K2 where K2≧

k by means of the coding function Cod2 of the second coder,

inputting the coded data words u²(1), . . . , u²(i−

1), u²′

(i) or u²(i), u²(i), . . . , u²(n) into the inputs of a second linear automaton circuit, which is described by the automaton equation

z²(t+1)=B·

z²(t)⊕

u²(t)

where z² represents a K2-dimensional state vector and B represents a K2×

K2 state matrix where B≠

A, and where the state matrix B can be inverted,

processing the coded data words u²(1), . . . , u²(i−

1), u²′

(i) or u²(i), u²(i), . . . , u²(n) by means of the second linear automaton circuit, the second linear automaton circuit,

undergoing transition to the state z²(n+1)=S₂(L2, y(1), . . . , y(i−

1), y(i), y(i+1), . . . , y(n)) if no error can be detected in the case of the data words u²(1), . . . , u²(i−

1), u²(i), u²(i), . . . , u²(n),

undergoing transition to the state z²′

(n+1)=S₂(L2, y(1), . . . , y(i−

1), y(i), y′

(i), y(i+1), . . . , y(n)) if an error is present at least in the case of the i-th position of the coded data words u²(1), . . . , u²(i−

1), u²′

(i), u²(i), . . . , u²(n),

the signature of an error-free data stream T_n being designated by S(L2, y(1), . . . , y(i−

1), y(i), y(i+1), . . . , y(n)) and the signature of a faulty data stream T_n being designated by S(L2, y(1), . . . , y(i−

1), y′

(i), . . . , y(n)),

determining the signature differences Δ

S1 and Δ

S2 by means of exclusive-OR logic combinations of the signatures S1 and S2 ascertained good signatures, in each case according to the following specifications;

Δ

S1=S(L1,y(1), . . . , y(i−

1), y(i),y(i+1), . . . , y(n))

⊕

S(L2,y(1), . . . , y(i−

1),y′

(i),y(i+1), . . . , y(n))

Δ

S2=S(L2,y(1), . . . , y(i−

1), y(i),y(i+1), . . . , y(n))

⊕

S(L2,y(1), . . . y(i−

1),y′

(i),y(i+1), . . . , y(n))

determining a unique solution for the position i of the faulty bit in the faulty data word by solving the equation

ƒ

₁<sup>−

1</sup>(A<sup>i−

n</sup>Δ

S1)=ƒ

₂<sup>−

1</sup>(B<sup>i−

n</sup>Δ

S2)

and if no unique solution results for 1≦

i≦

n, outputting a notification by means of an output medium that two or more errors are present in the data stream T_n under consideration,

determining a unique solution for the counter e(i) of the faulty data word y′

(i) in the data stream T_n by solving the equation

e(i)=ƒ

₁<sup>−

1</sup>(A<sup>i−

n</sup>Δ

S1)

outputting the position i of the faulty bit in the faulty data word and also the error e(i) of the faulty data word y′

(i) in the data stream T_n by means of an output medium; and

evaluating an integrated circuit in response to the output.