What happens if an isolation cell has the wrong clamp value, and how would this surface during verification?

Short Answer

The clamp value is what the always-on receiver sees while the source domain is off, so it must be the inactive level of that signal: 0 for an active-high enable, 1 for an active-low reset. Get it wrong and the receiver acts on a fake event as soon as isolation turns on, such as starting a DMA or resetting a block. Power-aware simulation shows it as a spurious event at shutdown, and static checks catch it only when the intended clamp is written down with set_port_attributes -clamp_value (UPF).

Technical Explanation

• Mechanism: while ISE is active, the isolation cell forces the crossing to 0 or 1, whatever the dead source does.
• Rule: clamp to the inactive level at the receiver. An AND-style cell clamps 0 and an OR-style cell clamps 1.
• Wrong clamp: the receiver sees an asserted signal for the whole off period, and edge-triggered logic fires once at entry.
• Simulation: the source goes X as its rail decays, the output jumps to the clamp value, and the receiver reacts right after ISE rises.
• Static: VC LP reports ISO_CLAMP_DEFAULT when a strategy has no clamp, and ISO_CLAMP_OVERRIDE when a port attribute clamp overrides the strategy clamp on that port.
• No tool knows your polarity unless you record it, so put set_port_attributes -clamp_value (UPF) on the ports that matter.
• What breaks: clamp 1 on an active-high dma_start starts a transfer every time PD_COP powers down.

# [UPF]  mychip.upf
set_isolation ISO_COP -domain PD_COP -applies_to outputs -isolation_supply SS_AON -clamp_value 0 -isolation_signal ISE -isolation_sense high
set_isolation ISO_COP_RST -domain PD_COP -elements {U_COP/rst_n_out} -isolation_supply SS_AON -clamp_value 1 -isolation_signal ISE -isolation_sense high
set_port_attributes -ports {U_COP/dma_start} -clamp_value 0
# [VC LP]  vc_static_shell
read_file -format verilog mychip.v
read_upf mychip.upf
check_lp -stage upf
report_violations -app LP

What To Check

• For every PD_COP output, the clamp equals the inactive level at the receiver.
• Active-low resets and enables get clamp 1 and map to an OR-style cell.
• The ports that matter carry set_port_attributes -clamp_value (UPF), so VC LP can compare.
• Power-aware simulation covers the shutdown entry, not only the steady off state.

Healthy, Suspicious & Hard-stop Results

• Healthy (illustrative): report_violations (VCLP) shows no ISO_CLAMP_OVERRIDE that changes a clamp value, and simulation holds dma_start at 0 for the whole off period.
• Suspicious (illustrative): ISO_CLAMP_DEFAULT on a strategy with no clamp: 0 was assumed, right for dma_start but wrong for rst_n_out.
• Hard stop: ISO_CLAMP_OVERRIDE on U_COP/dma_start: the strategy clamps 1 and the port attribute overrides it to 0, so the UPF disagrees with itself.

What The Interviewer Is Testing

• Whether you tie the clamp value to the receiver polarity, not to habit.
• Whether you know how simulation and static checks each expose a wrong clamp.

Follow-up Question & Model Response

"When would you choose a latch clamp instead of 0 or 1?"

Candidate Model Response: Use latch when the receiver must keep the last valid value the source drove, such as a mode bit read by always-on logic. The isolation cell captures the value when ISE rises and holds it through the off period. A fixed 0 or 1 would change that bit at every shutdown. The library must provide a latch-style isolation cell.