Homerun?
Problem 7Engineer
Lori designed and built a protocol converter using a Zilog Z80-compatible
Z182 communication processor combined with a 256K ý 8 flash memory
and a 512K ý 8 CMOS SRAM.
The prototype was tested for ESD hardness
by placing it on an aluminum plate and creating electrostatic discharges
of up to 10 KV at various points around the unit with a human body
model electrostatic discharge generator. Software was run during the
test which filled RAM with known values then continuously validated
memory contents while exercising the communication channels.
Lori found that the test software would
crash or fail randomly in one out of 10 discharges, and that the problem
occurred a bit more frequently at 5 KV than at 10 KV. RAM contents
were never altered by the failure. Which components or circuits should
Lori suspect first?
Answer:
The data bus would be the most
likely suspect since a disruption on the bus would only cause a problem
during the short time data is being transferred. The address bus or
control lines could also be a source of the problem. Lori should check
return paths through both ground and VCC, check for proper
bus termination at both ends and look for parasitic capacitive coupling
to the enclosure and mechanical components.
1-00
|