Your Functional Reliability Partner – IROC Technologies
+33 438 120 763 (FR)
+1 408 982 9993 (US)
info@iroctech.com
Contact Us
Support Site

TFIT Q&A

TFIT: QUESTIONS AND ANSWERS

1 What is TFIT?

TFIT predicts accurately and quickly the SER Failure in time(FIT) of any CMOS digital cell. Soft errors  is an issue not only in terms of design but also in terms of technology. TFIT is the only tool to integrate both dimensions , TFIT uses the cell SPICE netlist and layout as a design input and the transistor

TFIT allows  accurate calculation of the electrical effect of particles impact to a transistor , a cell, or a circuit early in the design flow, at much faster speeds than traditional 3D TCAD simulations (whereas the 3D TCAD approach does not apply in the case of neutrons impact and circuits analysis) . Furthermore, TFIT can be customized to the users’ technology just like internal TCAD effort, and integrate correlation with silicon test results. TFIT interfaces with Spice simulators so the electrical impact of the particle on a transistor is analyzed on a whole cell or circuit. Particles can be either neutrons (cosmic rays), alpha particles or heavy ions.

2 Why should I use TFIT?

TFIT addresses issues of different types of users:

  • Library developers. TFIT is a check for library developers using the latest technologies nodes who believe that Soft Error
    immunization is a business differentiator for their companies. TFIT gives the FIT rate for all the cells of the libraries. This is a data
    that more and more users of these libraries need
  • Designers who need to have a clear metric on how to measure rad-hardness. TFIT allows determining the LET threshold that
    triggers soft errors. In other words, the lower the particle LET (Linear Energy Transfer) that generates soft errors, the less immune
    the technology or design is. If this LET threshold is lower than 16, then soft error can occur for cosmic rays
  • Designers within fabless companies, who need to decide which foundry to choose for their next design at 90nm or 65nm, provided the models are available. This tool helps them figure out which foundry has the best performance for these technology nodes.

3 What are the input and output of TFIT?

The input data format is Spice netlist of the simulated circuit.

There are three types of output data:

Therefore there are two output files:

1- a text file with the results (FIT rate by nodes and transistors participation)

2- the spice netlist of the equivalent current generator (effect of the particle on the transistor) the voltage curves files (effect of the particle on the circuit).

4 Are location of impact and angle of impact taken into consideration in the model?

TFIT is designed for neutron impact, and as such, all the distributions of angles and impact locations are taken into account in the simulation through a simplified, but accurate geometric model. These distributions are the results of exhaustive analysis of all

combinations, not a Monte Carlo simulation. However, TIFT doesn’t automate the results of impacts for one single particle other than normal incidence (heavy ions case). We do support different incidences for single particles impact as a consulting service only.

5 What level of accuracy can I expect when calculating the standard FIT rate of a standard cell using TFIT?

TFIT simulation results are within the error bar of the radiation test results which typically range between 15% and 20%. So TFIT users can either take their own radiation test results or use TFIT simulations as an input to their circuit-level SER analysis.

6 What is a response model?

The TFIT response model characterizes a given foundry technology in terms of its sensitivity to radiation. It is a data base  of current pulses occurring in NMOS and PMOS transistors due to incident alpha particles or heavy ions or secondary particles created in neutrons reactions.

The TFIT response model is used in TFIT simulation software .It allows the designers to simulate the effect of radioactive particles on their design IP.

7 How do we get the TFIT response model?

Those models are generally distributed  to foundries customers as part of  the Process design kit(PDK) for a given technology. However  to be able to run soft error simulations, a TFIT software license must be acquired from iROC Technologies.

8 Which TFIT response models are available?

Available process models available so far:

TSMC 40G, 28HP.

GLOBALFOUNDRIES 40LP.

In development right now: TSMC 20, GF 28 and 20.

9 How does TFIT differs from other offers on the market?

TFIT is the only simulation tool to use the foundries’ process technology characterization for soft errors. Its accuracy and speed make TFIT a unique tool for soft error prevention and mitigation.