Cadence warning message about current exceeding 'imax' realistic importance

[pnsakanjankumar]

Friends
I am designing a circuit where I introduce explicitly high current exponential pulse which has a response time lesser than a CLK time period.
The warning message Cadence shows is "The bulk-drain junction current exceeds `imax'." So in reality if the current exceeds the limit is there a chance that it will lead to damage of the transistor even if the strike is for very less time.

This current pulse may occur due to radioactive decay emitting alpha particles

 

[erikl]

In a detector circuit configuration the drain current usually will be limited by an appropriate drain load - this still allows for energy resolution.

If - for any reason - you can't limit the drain current, a single event upset (SEU) won't damage the transistor, but a great lot of them, integrated, would ruin it - and not so much by exceeding the current limit, but much more by disturbing the lattice structure of the semiconductor: trapped α-particles with energies of several MeV may well displace atoms of the semiconductor lattice, by this increase the leakage current of bipolar devices, and change the threshold voltages of MOS devices.

Because of these effects, particle detectors are characterized for their (sort of) half-life (operation) time (or Vt-shift per dose, in case of MOS devices), depending on the type of particles, their energy and their capture cross section.

 

[pnsakanjankumar]

Thanks erikl
Actually I want to find the critical charge of the node. I am using double exponential current pulse and sweeping the Qcollected for which there is an upset (change in the state of the circuit).
I(t)=[Qcollected/(Tr-Tf)]*[exp(-t/Tr)-exp(-t/Tf)]
Let's say in a circuit, whatever the magnitude of particle strike you give the circuit state won't change (or state is restored) from the results of cadence (with the warnings). In that case what defines the critical charge of the node? is it the magnitude of the strike where there is a warning about the "current exceeding the limit imax" or when there is a oxide breakdown i.e.,
Warning - "The bulk-drain junction current exceeds 'imax'." occurs at Qcoll=50fC
Warning - "Oxide Breakdown" occurs at Qcoll=1000fC.

 

[erikl]

... what defines the critical charge of the node? is it the magnitude of the strike where there is a warning about the "current exceeding the limit imax" or when there is a oxide breakdown i.e.,
Warning - "The bulk-drain junction current exceeds 'imax'." occurs at Qcoll=50fC
Warning - "Oxide Breakdown" occurs at Qcoll=1000fC.

I don't understand this 1^st warning: "The bulk-drain junction current exceeds 'imax'." occurs at Qcoll=50fC
Say the collected current pulse has a duration of only t=1ns, then I=Q/t = 50e-15As/1e-9s = 50µA -- not really a high current, nothing which could destroy anything.

An estimation for the 2^nd warning: To generate one electron charge in silicon, you need an energy of the order of the band gap energy, 1.2eV in silicon, let's calculate with 1eV : A single α-particle with an energy of, say, 5MeV , transferring its total energy and creating 1 electron charge per eV would create a charge of 5e6 *1.6e-19 As = 8e-13 As =800fC . In a junction (either depletion zone or gate-to-channel) of C=100fF this creates a voltage V = Q/C = 800fC/100fF = 8V .

This is for a (simultaneously) single impinging 5MeV particle. Hence a protection diode between gate and bulk is advisable, same as for antenna protection.

 

[dick_freebird]

These "assert" statements in my experience can have zero relation to any
known limits. They are often left as defaults by lazy modeling engineers.
However it behooves you to figure out whether or not there is any real
value behind the limit, and whether the "violation" is any kind of physically
realistic outcome or just simulator garbage.

Your single event charge collection dynamics if unrealistic could induce
this kind of current, just in the junction capacitances. The simulator does
not care whether the "overcurrent" is for 1fS or 1 hour.