Re: USB on a breadboard?
- From: "Noway2" <no_spam_me2@xxxxxxxxxxx>
- Date: 8 Aug 2006 06:24:08 -0700
brehob@xxxxxxxxx wrote:
Hello all,
I'm at instructor at a large US school who has slowly been moving into
teaching embedded systems. I know very little about embedded systems
in the real world (I'm more of an architect), but I've learned a lot by
teaching and working with students on embedded projects.
We currently have a class that uses a combination of an MPC823 and an
FPGA to do software/hardware design. It is mostly in assembly (a bit
of C) and covers the details of doing memory-mapped I/O, designing the
hardware devices, interrupts (in gory detail), timers and A/D stuff.
I've been trying to upgrade our embedded systems offerings and so I am
offering a class next semester that is a follow-on to the current one.
I've have chosen to go with two different platforms, an ARM based
system from Intel (sitsang) with an OS, color touch LCD, and crazy
amounts of I/O as well as an Atmel based portion (ATmega8) where the
students will design (and get built) a board.
So now the questions:
#1 Is there a USB chip that I could possibly talk to with the ATmega8?
I'd love something I can put on a breadboard without having to solder
very small wires. Looking at the student projects I'm expecting, USB
may be a pretty common desire. The sitsang board does USB in its
sleep, but I'd like them designing their own hardware where possible.
#2 Same as #1 but for ethernet.
#3 What do _you_ all think is important for them to know? If you were
to hire one of our graduates, what specific skills would you be looking
for? Are skills like being good with solder important to any/many of
you? Just software?
#4 What am I missing?
More background:
I'm going to teach a fair bit about how to interface with an OS (for
the sitsang) and the basics of Linux drivers. I'm also going to spend
a short time on the AVR architecture and interupts (which I'm still
learning) as well as a bit on the basics of board design. I'm planning
on having a few guest speakers come in and talk about what they do (one
a WIMS person, one works for Rabbit, and maybe a Lockheed-Martin
embedded person).
The major goal of the class is to give them the oppertunity to design a
system (probably involving board design) of their choosing. We have a
reasonable budget (thanks Lockheed-Martin!) of around $800 per group of
4 students. So we should be able to build some cool things!
Answers to the above or random thoughts greatfully accepted!
The first thing that I would like to say is that I am pleased to see
that these types of classes are being offered and I am impressed at
your initiative.
In response to your questions, hardware development with USB and
Ethernet may be a little bit tricky, from the breadboard perspective.
The reason being is that both technologies utilize differential
signaling on lines with controlled impedances which would be difficult
to achieve with a breadboard. I think the suggestions you received
recommending a development kit or adapter are probably the way to go.
This way you could at least focus on the concepts of the protocols and
possibly hook up a scope or something to look at the electrical
signals.
Second, I think that your focus on interrupt handling and uP
interfacing is very good. As an EE who was worked with embedded
systems for the 10+ years since graduation, I find that these skills
have come in handy on just about every project. Based on my experience
with a computer science person who reports to me, I think these skills
could have used some more attention in his education as they made his
eyes glaze over when he had to deal with low level issues that would
normally be handled for you by an OS. Generally speaking, his ability
to program software was excellent, but he would become befuddled when
faced with hardware interfacing. Granted, he is a CS major, but not
everybody who graduates for the program will go work for Microsoft.
Third, your work to interface with Linux is an excellent idea, for two
reasons. First, there is the aspect of embedded systems that focuses
on using single board computers (PCs) with a scaled down operating
system, such as micro-Linux. Second, it is sometimes possible to
obtain a development card that interfaces to a PC to do some early work
prior to the creation of the target application. As Linux allows a
user to modify the kernel, and place handlers and other types of code
in the system it makes an excellent platform for this type of
development. Windoze (spelling deliberate) isn't nearly as flexible as
with it, you are the tail and the OS is the dog.
I also think that your idea about some soldering experience has merit.
I believe that being able to perform some basic board rework is a
valuable skill for both engineers and technicians. A course that
provides IPC certification takes about two days so it should be
possible to cover a fair amount of technique in a few class or lab
sessions. I personally, populate all my new designs on the first
prototpe. This allows me to populate sections like the power supply
and reset circuit to verify it before populating the rest of the board.
While it may be an unconventional approach, it certainly makes me more
comfortable than the idea of throwing the switch and hoping that my
last four months of effort doesn't go up in smoke.
I think that introducing FPGA design is an excellent idea too. Not
only are FPGAs becoming increasing common in many designs, I believe
that being able to work with them has helped my career based on
employer comments in annual reviews.
You may want to consider introducing some of the aspects of real time
design and programming. Real time issues are inherent to some degree
in just about every embedded system. In contrast, though, in most PC
based applications, the user has little control over the timing. I
think uCOS-II might be an excellent avenue to pursue this.
Other areas to consider discussing would be the importance of system
reliability and how this is achieved through the use of watchdog
circuits and proper coding practices. Also you may want to discuss the
importance of good user interface design and how often times the
designers think very differently from those who use the systems.
The only other thing I can think of that would be helpfull, that you
didn't explicitly mention is some experience with development tools
like how to properly use oscilloscopes and logic analyzers and most
importantly, understanding their limitiations.
.
- Follow-Ups:
- Re: USB on a breadboard?
- From: brehob
- Re: USB on a breadboard?
- References:
- USB on a breadboard?
- From: brehob
- USB on a breadboard?
- Prev by Date: Problem with SDRAM Memory
- Next by Date: Re: Can you be a good programmer?
- Previous by thread: Re: USB on a breadboard?
- Next by thread: Re: USB on a breadboard?
- Index(es):
Relevant Pages
|
