Boundary Scan over In-Circuit Test?
It's often heard and read that In-Circuit Test has come to its limits in terms of board accessibility and
utilization flexibility. Is Boundary Scan the logical successor or what makes most sense for achieving
highest possible test coverage?
The question is far to simplified. There are numerous factors that determine method of testability.
With regard to Boundary Scan:
1. Are there boundary scan devices on the board?
2. Is there a chain of boundary scan devices or only one?
3. How do I power the board?
4. Is there connector access to the JTAG signals or do I need a fixtured access?
5. Are there mechanical needs such as heat sinks that must be attached if power is applied?
Any of the above could remove Boundary Scan as a viable test method and force the issue to ICT.
But then the limitations of fixtured access come into play. The combination of Boundary Scan and
ICT is a powerful tool and can reduce the mechanical access that normal ICT requires. Too little
fixtured access and the ICT process even with boundary scan are not cost effective.
There is another process here that is ignored. The combination of Flying Probe Test and Boundary
Scan. The limits of Boundary Scan are with the analog and non-boundary scan devices. The limit of
flying probe is the ability to provide sufficient stimulus and measurement to make powered testing
viable. When Flying Probe and Boundary Scan can be combined many of the limitations for the
stand alone tests are removed. Full Boundary Scan coverage, full flying probe coverage and the
extended coverage that using the flying probes as mobile stem and response units for the boundary
scan. The NRE for a test fixture is gone but due to the extended test time the per-unit test cost
No one process can fit all types of board, the need is to determine best available method and the
starting point is always the design of the board.
Sr. Test Engineer
There is no right or wrong answer, perhaps just a more correct answer.
In-Circuit Test is a very powerful tool for testing printed circuit boards, below is a short list of what
can be tested for:
Miss oriented devices
Out of tolerance device values
As circuit boards have become more electrically complex and physical component densities have
increased to unimaginable levels ICT has also evolved and become much more capable. As stated
with other’s answers to the ICT vs. B Scan; the board and devices must be boundary scan compliant
and you must have access to their test ports. Please remember ICT testers also have different
capabilities, and a thorough well thought out test plan, test program and fixturing can make a world
of difference in gaining acceptable test coverage versus an average test. As with many of today’s
advanced telecom boards’ electrical networks far exceed the available resources of today’s best ICT
testers. This doesn’t render ICT useless by any means, advanced testing uses several different
methods which have resolution down to the component lead level, or in today’s terms BGA ball. In
many cases today’s ICT testers roll Boundary Scan in to their test tools to enhance test total
Just as ICT usually is not the sole test for a circuit board, it is just one of the tools required to
complete the coverage. Total coverage may require automated optical inspection (AOI), automated
X-Ray inspection (AXI) and in some case flying probe testing followed by a functional and system
test. Boundary Scan rolled into the ICT test is just one of many tools required to test today complex
board. ICT does provide the resolution required to easily repair the printed circuit board before
incurring any additional value add.
Circuit Check, Inc.
Boundary-scan plays a large role in solving the problem of diminishing test access from ICT.
Combining it with other non-intrusive board test (NBT) technologies such as processor-controlled
test, FPGA-controlled test, IJTAG test, and High-Speed IO testing will take you a long way in the
quest to counter the diminishing test coverage on expensive heavy irons like ICT, and Flying Probe.
So the question “Would you recommend non-intrusive board test (NBT) over In-Circuit Test (ICT)?”
would definitely be YES, providing the design allows test NBT type testing.
One tip is to start with a potent, highly automated boundary-scan tool as a base, and add other nonintrusive test technologies to counter the diminishing test coverage on expensive heavy irons like
ICT, and Flying Probe.
Here’s some interesting reading: Solving the Problem of Diminishing Test Coverage from In-Circuit
Kent Zetterberg ASSET Inter Tech Inc
ASSET InterTech Inc.
It took a while before the first (or the second) Boundary Scan vendor reacted to this question :-)
As the opener of this thread I'd like to refer to GOEPEL electronic's JTAG/Boundary Scan
solutions, being far more than just tools for test coverage increase or test problem track down. They
help all the way for test, programming, validation, emulation or debugging:
In this day and age of highly complex electronics and firmware, I would recommend doing
functional testing of boards.
For some reason this question shows up on the daily e-mail (at least the one I received today) even
though it was asked several months ago.
To determine the proper test strategy you need to ask why you test. From my perspective, there are
two philosophies to testing, one to detect defects and one to verify performance to specifications.
Depending on which philosophy is chosen, or a combination of both, will determine the test strategy.
If you are testing to detect defects, perhaps the best way to begin is with an FMEA – Failure Modes
and Effects Analysis. What possible defects could occur, what is the frequency, likelihood and
severity of these defects, and what can be done to detect them? Decide what is an acceptable amount
of risk in having a defect escape, and design your test strategy and process accordingly. Your
strategy may include ICT, functional testing, boundary scan, AOI, visual inspection, x-ray and the
like in order to accomplish your goals.
If you test to verify performance to specifications, you should also analyze the test process to
determine if possible defects can escape and what effect this may have on your product in service,
for example, can reliability and durability be compromised?
Then, to top it all off, you must look into the cost and throughput of your test strategy and determine
if this fits in with the overall process parameters for production to meet demand and profitability.