The procedures below used SDR (Software Defined Radio) as a case study which summarized the testing process & ideas through the whole design process as a reference for future design.
When subsystems are assembled / compiled, they are ready for unit testing.
Unit testing determines if you satisfied the requirements as set out by ICD and the design proposal
provide power to interconnect all subsystems
has connectors mates with each subsystems
breaks out relevant signals to connectors (eg. BNC) on the edge of the board
allows both individual test of each subsystems and integration test
provide easy connection to test equipment
polarity connection
be careful about positive & negative terminals to avoid backword connection to damaged & rendered usability of main board
install one PCB at once
this depends on the testing plan (some plan settings and test board assume all other PCBs are absent)
set current limit on power supply
Normally, a current limit of 100 ~ 200 mA is enough
before powering up, set a constant voltage & a current limit on PC, then turn on the switch
correct output ✅ : PC stays in CV mode, current shows a small & expected value (50 mA)
wrong output 🚨: PC immediately switches to CC mode, huge voltage drop, current hits the limit, smells hot
-> solution: shut it off right away, look for shorted power rails, any wrong orientation, measure VCC ~ GND resistance
Need for Dummy Load?
made by precision resistor & heat sink, important to consider when test power amplifier & RF circuit individually
electrically acts like an antenna (real load) but doesn't radiate
purpose:
impedance matching -> avoid signal reflection, stress damage, high voltage standing wave ration (VSWR)
actively pushing circuit current (power) into the dummy load as an output
prevents thermal runaway (unintended RF radition) or oscillation ( unstable operation & teminatio)
Sample Usage of Dummy Load
Documentation:
It will be helpful to plan following aspects before doing unit testing:
list of test equipment (including options), procedure (notes before begin, detailed steps), questions to consider according to ICD, list of measurements for data collection
could follow right document as an example
Collection of Usable Procedures:
Testing without switch installed:
use jump wite, hard-wire the output, observe the signal there
Set output termination:
For CMOS & Oscilloscope, use High - Z -> so function generator assumes it's driving a high-impedance load, correctly deliever and display the intended signal
otherwise: wrong load, wrong output voltage, oscilloscope may display double the expected amplitude
when signal frequency > = 10 MHz:
select "sine wave" at "waveform" , the function generator cannot generate non-sinusoidal signals with a frequency higher than 10 MHz -> even if I want a square wave
when systems exist multi-input signals (eg. 3.3v, 5v):
identify the function of siganls (3.3v controls high & low stage, 5v is power signal), and what should respond
first, set 3.3v high & disable other parts of circuit (eg. PA), discover current draw + set 3.3v low, repeat
enable function generator observe the output -> calculate power
total harmonic distoration (THD):
use FFT by automation scripts
measure the amplitudes of various harmonics (n >= 5) for various output power