Coherent syndrome extraction for the three-qubit repetition code ($(s_4,s_5)=(b_2\oplus b_3,b_1\oplus b_3)$)

Encode $u|0⟩+v|1⟩$ as $u|000⟩+v|111⟩$. Four controlled-NOT gates copy the displayed two data parities into initially zero quantum ancillas. No logical amplitude is measured. The error syndromes $00,01,10,11$ correspond respectively to no bit flip and bit flips on data qubits $1,2,3$. Conditional Toffoli gates, with temporary Pauli X gates implementing zero-valued controls, undo the indicated error. Inverse encoding yields the original logical state and a separate syndrome state. A phase flip has zero bit-flip syndrome and remains as a logical phase error; this three-qubit code is not a general single-qubit-error code.

Suppose the first ancilla records parity $q_2\oplus q_3$, then a fault of value $f$ flips $q_2$, and the second ancilla records parity $q_1\oplus q_2$. For incoming bit-error pattern $e$, the observed error syndrome is the displayed pair, while the final data error before recovery is $d=e\oplus (0,f,0)$. When $f=1$, the observed syndrome differs from the actual data syndrome by $(1,0)$. Applying the usual minimum-weight recovery leaves a nonzero syndrome, so the state is outside the code space even when the channel itself made no error. This is a timing-dependent failure of this simple extraction circuit, not a general obstruction to fault-tolerant quantum error correction.

For independent channel bit flips of probability $p$ and the specified extraction fault of probability $q$, require recovery of an arbitrary input in the original encoded subspace. Every extraction-fault event leaves nonzero error syndrome; without that fault, channel errors are corrected exactly at weight zero or one. Thus success probability is $(1-q)[(1-p{)}^3+3p(1-p{)}^2]$. Compared with a single unencoded bit-flip channel, strict improvement for $0<p<1/2$ requires $q<p(1-2p)/[(1-p)(1+2p)]$. A different decoder or an added later ideal recovery changes the performance criterion and must be specified separately.