Milestones and Debug Techniques for ECE 189B Spring Quarter Projects, Study Guides, Projects, Research of Electrical and Electronics Engineering

Download Milestones and Debug Techniques for ECE 189B Spring Quarter Projects and more Study Guides, Projects, Research Electrical and Electronics Engineering in PDF only on Docsity!

• Spring QuarterMilestones & DebugTechniques ECE 189B --- Mar 31,

Boards are now available(!)
This year’s PCB fab andassembly is complete^ ^ Group leaders: please stop bymy office to pick up your group’sboards

BluKey

SureStop

More Project Boards …

T-

ThermIN

Milestones for Spring Qtr ‘09^
^ MS#1 ---- Initial Power-Up^
^ MS#2 ---- Processor / Memory / Boot-up^
^ MS#3 ---- BIOS-Level Monitor or O/S^
^ MS#4 ---- Individual Subsystem Tests^
^ MS#5 ---- Integration of Subsystems^
^ MS#6 ---- Full Application^
^ MS#7 ---- Notebook, Poster, & Presentation

Presentation Day is Thursday, June 4

Before Milestone #1
Update your documentation^ ^ For many groups, the schematic was changed after

the notebook was

turned in  Print out a new, up-to-date version and file the old one(s)  Make sure^ your schematics and write-ups match what was actually built!^
^ Note that for some projects the assembler may have done some “creative”things! ….^
^ Make sure you understand what the assembler did

and did not^ do

^ You should pick up the notebook turned in at the end of fall quarter. Usethat to build upon …
Think about the current state of the board ^ some projects are still a bit incomplete ... ^ figure out what components are needed for MS#1^
^ Do^ not apply^ any^ power

until you have thought about it carefully

The Moment of Truth … Power-Up
First rule: do not

trust anything

^ The circuit you are about to test has never been tried ^ You don’t know anything about its correctness ^ Treat the whole event as an experiment^
^ assume the worst, but be prepared to be pleasantly surprised^
^ be aware and react immediately to high temperature or smoke(!)or unintended light(!)^ Remember that once you let the factory-installed smoke out of acomponent …. you have to replace it before proceeding

More about Initial Power-Up
Make sure you know where to measure; work quickly ^ We don’t want things heating up while you are finding your probes,eating a sandwich, going to your locker, or turning on the meter ^ It may be necessary to hold your processor(s) in reset^
^ Why?
Once the board can support the full supply voltage … ^ Measure current drawn from the power supply … log it (per board)^
^ Keeping a log should be part of your MS#1 write-up for the notebook ^ Measure & check all derived voltages (e.g. 9V

5V, 5V^

3.3V)

^ If possible, measure the current drawn from each ...  Monitor the heat in all ICs
^ Make sure nothing is heating up
^ If one or more chips are getting hot, turn off the power and get helpfrom TA or instructor

Milestone #1 … Power-Up
^ Ultimate goal: board is powered up with thevoltage(s) as planned and with currents that arewithin specification^ ^ Milestone #1 is achieved when your board has earned^ your^ trust^
^ i.e. you’d be willing to leave the board powered on all nightwithout supervision or checking … knowing that everythingwould be OK. (Don’t try it ... unless you know it’s OK)^ ^ Once one board has achieved MS#1, attempt to repeatfor the other boards (but not the saved virgin board)^
^ Keep a log of the board numbers and measured current drawn^
^ If mods must be made, keep a log of the status of each board^ 

Extremely important … do

not^ cut^ corners

on patch documentation

Program DownloadYour mileage may vary based on processor &interfaceSome systems use JTAG, some UARTs, somehave flash
JTAG: wiggler and special cables needed($$!) ^ Must use in association with an IDE thatsupports it ^ Excellent debug environment, once JTAG^ cabling & SW figured out

Prototype Debug Techniques
^ First-time programming of CPLDs, FPGAs^ ^ It may be necessary to hold reset condition activewhile programming^ ^ Issue? Programmable logic may be generatingcontrol signals that conflict. It is likely that CPLD orFPGA outputs will float

while being programmed

^ Software boot-up^ ^ Prepare a simple

sign-of-life program ^ Initialize processor, disable interrupt system, wigglean I/O line or flash an LED or write “Hello, World!” to aUART channel.^
^ Get the cabling built and checked out

Patching Hardware Errors

(continued)

^ Deletions

:^ ^ connection(s) need to be removed

^ Be careful … hardware mods aren’t very undoable ^ If you can see the PCB trace, cut it with an Exacto knife ^ If trace is completely under an SMT component, youhave 2 choices:^
^ Lift the lead off the pad (can be quite difficult for fine pitch SMT)^
^ Clip the lead (ditto) ^ If all else fails …^
^ get help to remove entire SMT component with “low-melt”^ ^ avoid using temperatures that are too high …. Be careful!^
^ modify board or clip the leads on a new component^
^ re-install …. can be tedious

Adding Components
Often difficult to add

new components

^ No place to mount them, no physical attachment ^ Sometimes you can find an existing sub-circuit that canbe sacrificed or re-targeted for use ^ For discrete components (R or C) added in

parallel

^ stacking is sometimes possible for 0805, 1206, etc.  Look for vias or thru-hole components for anchoring thenew component(s) …… plan, plan, plan

Subsystem Replacement
Sometimes things are simply too messed up to berepairable^ ^ Build a little perf-board with a new subsystem on it (hand-made)and attach somewhere convenient^ ^ Select the subsystem interface carefully to cut down the number oftraces that must be accessed^ ^ e.g. if it is a regulator, disable or remove old one^
^ Build new sub-board and attach it to battery subsystem

Documentation
Always^ keep track of any changes made to board(s) ^ Label each board with a unique

number

^ Keep status log showing what has been done to each ^ It is a good idea to always keep at least one -- more, ifpossible -- virgin (unmodified) boards
Project status file or website or notebook is a greataid for improving group communication, e.g. ^ “I worked over the weekend and got XXXX to work by patching U101,pin 12 high on board #2. ---- Joe”