SELF-CHECK / SRAM WRITE BOUNDARY
Solve from the evidence.
Do not guess the plot.
SRAM Write Boundary Self-Check uses answer reveal blocks only. It does not store scores, identify visitors, or collect submissions.
QUESTION SET
Eight checks before reuse.
Reveal an answer only after writing your own. The goal is to catch overclaims before they enter a lab note, portfolio, or research discussion.
01 / conceptState the initial cell state and intended write direction.
The initial state is Q=0 and QB=1. The intended write flips the cell to Q=1 and QB=0.
02 / evidenceWhat are the two sweep axes in the shmoo?
The axes are wordline amplitude and cell VDD during the 3 ns write pulse.
03 / metricWhat exact pass condition decides whether a case succeeds?
At 7.5 ns after supply restoration, Q must be greater than 1.35 V and QB must be less than 0.45 V.
04 / interpretationAt cell VDD assist 1.4 V, what minimum wordline voltage succeeds?
1.0 V is the minimum successful wordline voltage for the 1.4 V assist condition.
05 / comparisonWhy does stronger VDD collapse move the boundary?
Lowering cell VDD weakens the cross-coupled latch during the pulse, so the access path can flip the cell with less WL drive.
06 / boundaryName two things this result does not claim.
It does not claim measured silicon behavior, and it does not include mismatch, full PVT, extracted parasitics, half-selected cells, sense amplifier behavior, or driver impedance.
07 / downloadWhich public files let a reader inspect or verify the data?
The JSON payload, summary CSV, generated netlists, run script, verify script, reproduce wrapper, and SOURCE_NOTES_WRITE.md are published under /data.
08 / portfolioWrite one sentence a learner could use in a portfolio without overstating the result.
Example: I reproduced a SKY130A TT pre-layout 6T SRAM dynamic write-success shmoo across 24 transient cases and documented the pass/fail boundary and limitations.