A holding circuit enables parallel repair-circuit selection, preserving memory redundancy repair paths when circuits or wiring fail.
A mode register flags invalid memory writes so hosts can retransmit data or pause applications after command faults.
A repair circuit switches between shared and individual wordline repair to match failed-cell patterns across memory banks.
Transmission and reception counters plus voltage-gated error signaling help detect invalid memory data while limiting power use.
ECC-coded reference data guides read-voltage updates, adapting NAND flash reads to charge loss and inter-cell interference.
This case merges staircase and CMOS contact etches in 3D NAND to cut wafer cost and avoid liner damage.
Stacked cell and peripheral structures use vertical channels and interconnects to increase density, speed operations, and reduce leakage.
Vertical memory-cell integration uses through-layer vias to limit parasitic capacitance.
Dual-edge sampling and pre-delay align DRAM commands for accurate 1N operation.
Periodic RDCL scans target high-risk word-line pages and fold error-heavy blocks after high-temperature read stress.
On-die processing circuitry tests memory internally, using faster clocks and analog-to-digital conversion for finer failure data.
A register, comparator, inverter, and controller support parallel memory testing with low time overhead while preserving stored data.