Can you describe a basic low-power implementation mistake, like a missing isolation cell?
From PDVerse Low-Power Physical Design Mentor Guide, part of the pdVerse Mentor Guide
Short Answer
A missing isolation cell leaves an output of a switchable domain driving always-on logic with nothing to hold it when the source powers down. The receiver sees a floating value, which causes crowbar current and random logic behaviour only when the domain actually turns off.
Technical Explanation
- Symptom: power-aware simulation shows X on the receiver after shutdown, or silicon draws extra current and misbehaves in sleep.
- Normal functional simulation and DRC/LVS pass, because they never model the source domain being off.
- Detection:
check_mv_design(ICC2) reports the missing cell;check_lp(VCLP) reports ISO_STRATEGY_MISSING at the UPF stage and ISO_INST_MISSING at the design stage. - Cause: usually no
set_isolation(UPF) covers that port, or its filters exclude it. - Fix: write or widen the strategy with an always-on
-isolation_supply, then reruncreate_mv_cells(ICC2) and the checks. - Legacy form (still accepted by ICC2/PT):
-isolation_power_net/-isolation_ground_netonset_isolation(UPF), plusset_isolation_control(UPF) for the signal.
# [UPF] design.upf
set_isolation ISO_COP_OUT -domain PD_COP -applies_to outputs -isolation_supply SS_AON -clamp_value 0 -isolation_signal U_PC/ISE -isolation_sense high -location parent
# [ICC2] icc2_shell
create_mv_cells
check_mv_design -isolationWhat To Check
- Every output of a switchable domain whose receiver can stay on is covered by a strategy.
- The isolation supply is always-on.
- The clamp value suits the receiver: 0 for active-high inputs.
- The isolation signal comes from always-on logic.
Command Checks & Actions
set_isolation ISO_COP_OUT -domain PD_COP -applies_to outputsCover every PD_COP output with isolation
check_mv_design -isolationReport crossings still missing isolation
report_mv_path -cell U_COP/DATA_UPF_ISOShow why a cell was or was not inserted
check_lp -stage designStatic check of the netlist against the UPF
report_violations -app LPList isolation violations by tag
Healthy, Suspicious & Hard-stop Results
- Healthy (illustrative): No isolation errors in
check_mv_designand no ISO_* errors in check_lp. - Suspicious (illustrative): Isolation present but clamp or supply unchecked on a new port.
- Hard stop: An output of PD_COP reaches PD_MYCHIP with no isolation cell.
Common Mistake
The Trap: Skipping isolation because the two domains run at the same voltage.
- Isolation depends on whether the source can be off while the receiver is on, not on the voltage. Same-voltage domains still need it.
What The Interviewer Is Testing
- Can you walk from symptom to detection to fix?
- Do you know why normal simulation misses it?
Follow-up Question & Model Response
"Why must the isolation signal come from always-on logic?"
Candidate Model Response: The isolation cell has to hold its clamp for the whole time the source domain is off. If its enable came from the switched domain, the enable would float at the same moment the data does. The clamp would then release at random. So the power controller in the always-on domain drives it, through always-on buffers if it needs buffering.
Practical Example
Design Scenario: (illustrative) In MYCHIP, U_COP output irq_out feeds an always-on interrupt flop in PD_MYCHIP. Both run at 1.0 V, so the designer skipped it. When PD_COP turns off, irq_out floats and fires false interrupts. Adding ISO_COP_OUT with an AND clamp to 0 (ISO_AND_X1, an illustrative cell name), enabled by U_PC/ISE, fixes it.
Low-Power & UPF Handbook
Master Low-Power VLSI & Multivoltage Design
Read the complete low-power guide library covering power domains, level shifters, isolation clamps, state retention, and UPF signoff verification.
Offline PDF Bundle
Want all 1109 questions offline?
Get the complete 4-book PDF bundle (PnR, STA, MMMC, Low Power) with a clickable table of contents - no ads, no internet needed.

Continue practising