140 likes | 289 Views
analogue embedded testing. The question is?. Analog & Mixed-Signal Test Where are we going to?. or The question should be:. Analog & Mixed-Signal Test Where do we have to go?. The answer is:. We are going to where we have to go. That is. On-Chip Test Manager. Stimuli generation
E N D
analogue embedded testing The question is? Analog & Mixed-Signal Test Where are we going to?
or The question should be: Analog & Mixed-Signal Test Where do we have to go?
The answer is: We are going to where we have to go ... That is ....
On-ChipTest Manager • Stimuli generation • Result compression • Precision timing • Diagnostic • Power manager • Test Control • Support for board & system level DfT & BISTed ICs Low Bandwidth external interfaces DISTRIBUTED or CONCENTRATED? Memory (BISTed) External ATE Digital Tester Low cost-per-pin Limited speed Limited accuracy Logic (BISTed) Mixed-Signal (BISTed) I/O & Interconnects (BISTed) IC High Bandwidth internal interfaces
The SUPER-COSTLY-TESTER Era The ATE industry is creating SUPER-COSTLY-TESTERs as a demand of the IC manufacturers. Memory: • Very high amount of data. Represents 40% of improvements in testers Mixed-Signal Instrumentation: • Higher bandwidth, higher sampling rates, higher accuracy, lower noise, etc. • RF and audio circuits a major challenge, more when noisy digital circuitry is also present. DUT to ATE interface: • Higher pin-counts, high frequency & performance probes and sockets. • No degradation of tester accuracy and noise.
Test cost Is in fact a money problem? May be. Perhaps not ! However …
Source: SIA Roadmap uP Memory digital Devices30% per year RF MixedSignal ATE: +12% per year aprox. Tester timing errors Bandwidth Gap between ATE and signals On-Chip!!!!!
Test Complexity More to Test but Less Test Access!!!!! Test Development and Application time may become prohibitive
The KEY AREAS • Standard Test Access Mechanism • 1149.4 well-accepted • P1500 needs MS extension • Structured Test Planning: • Depends on the available resources (DSP, uP, Memory, etc.). • DfT & BIST
DfT & BIST • Can be emulated on-chip a complete external testing? • NO for most of the functional testing cases.SO, • Functional testing must be reduced. • Reduce costly or non-practical on-chip tests • Structural test is a clear powerful complement to functional testing. • Structural test related to the I/O functional behavior and performance are of main interest.
CONCLUSIONS MORE RESEARCH IN THE KEY AREAS IS NEEDED!!!! So, Is in fact a money problem? May be. Perhaps not? YES!!!!! Because Research can not be disaggregated from money.
There is no other way to cope with the cost and complexity of future ICs. • Do not try to emulate a complete external testing: This is not practical in most cases. Functional testing must be reduced. • Reduce costly or non-practical on-chip tests • Structural test is a clear powerful complement to functional testing (OBIST) • Structural test related to the I/O functional behavior • Accessing: Yes, but non-intrusive (sw-opamp)
KEY AREAS • Re-usable and structured DfT & BIST techniques • Reduce I/O data rate requirements, • Enable low pin count testing, and • Reduce the dependence on expensive instruments. • Structured Test planning • Enable hierarchical testing • Enable the re-use of on-chip resources (DSP, uP, etc.) • Facilitate parallel testing • etc. • Standardized Test Access Mechanism