Re: Suggested Starting Point
- From: paul$@pcserviceselectronics.co.uk (Paul Carpenter)
- Date: Tue, 04 Jul 2006 11:34:52 +0100 (BST)
On 3 Jul, in article
<1151987686.620715.174570@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
deja@xxxxxxxxxxxxxx wrote:
Thank you for taking the time to respond.
Paul Carpenter wrote:
On 3 Jul, in article
<1151948309.418816.240620@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
deja@xxxxxxxxxxxxxx wrote:
Hello,
What I *think* I want is a single-board computer that:
*Can fit in a 4"x4"x1" space (can be bigger, but not much)
*Has 1-3 USB ports that can handle a web cam, wireless ethernet, and a
USB drive (thumb or SDcard) from which it can boot.
Why have wireless ethernet from USB? Sounds like you will need extra
drivers and layers of software to gobble up RAM and overhead slowing
your system down and making the main board _possibly_ more expensive.
Due to extra USB and RAM requirements, when they may well have ethernet
ports already and wireless access points are very cheap anyway.
You're right. USB because I knew that many SBC computers have them
(even my old 133mhz Pentium MMX laptop has a USB port), and I know that
USB 802.11b/g adapters are small and cheap. Most PC104 boards do have
ethernet as you indicated - though few (that I've found) are wireless.
Obviously if I can get it onboard then I would not also need to go
through the USB port. Seems like even if I go the PCMCIA route that it
would still require drivers of some sort.
Consider this ethernet port or wireless adapters appear as network adapters
to the OS so the data is sent from/to network card as packets to/from the
network software stack (in simplistic terms).
For USB to wireless/ethernet adapters the data is sent from/to the network
stack to/from a pseudo network driver, that then encapsulates the data
to send to the USB controller driver (under the same OS) to send to the
USB device to de-encapsulate the network traffic to send to/from the
wireless/ethernet.
The extra hoops and encapsulation steps (let alone doing any compression)
puts extra load on the cpu and RAM requirements, which in turn limits your
dataflow. Also more layers of software/drivers to go wrong, be incompatible
and other such problems.
Even a PCMCIA wireless card is closer to a network card than a USB device.
*Preferrably is compatible with the x86 architecture
*Would like to have a parallel printer port
*Has sufficient Memory - (Not really sure at this point : enough to run
apache/php in addition to the OS (Linux))
*Speed-(Not really sure - see "the jist of it")
*Is Cheap (less than $100 - preferably less than $50)
Considering your requirements and later items I doubt you will get
near this price.
About the best I've found so far is $69 for a new PC104. I could go
that high. I'm also contemplating an HP 720LX handheld which are going
for about $35 in a "used" condition.
I would still suggest using a dedciated USB web cam server plus wireless
access point and a LARGE battery.
e.g. <http://www.digidave.co.uk/product_info.php?products_id=254>
Plug your USB 2.0 camera into the server, test with wired ethernet first.
(this unit can take TWO webcams).
The server is a small box that costs about 80.00 UK pounds!
If you were willing to replace the camera there are webcams like ones
reviewed here that are webcams to 802.11g in one unit around 100 UK pounds
already.
<http://www.filesaveas.com/wifiwebcam.html?gclid=CLmBgLH194UCFSFsEAodyXIjuA>
Even the Dlink wifi camera DCS-5300W can be got for around 70 UK pounds.
That is your camera and most of your system.
Alternatively look at
<http://www.x10.com/key_products/wireless-webcam.htm>
This is wireless but to a TV not the web.
Looking at wireless CCTV that gives a receiver and TV output such as
<http://www.y3kvision.com/acatalog/3k15.html>
About 50 UK pounds including camera and receiver.
These from simple google searches for 'webcam server' or 'wireless
webcam server'.
I don't need keyboard input, video, or HD/FD controllers
The jist of it:
I would like to be able to run a wireless webserver that can serve up
web-cam pictures at 30fps.
Have you done the calculations for how much data just the 30fps at
whatever resolution you are choosing will be?
Have you even worked out what picture resolution you need or want?
I doubt it to BOTH of the above questions!
You're right. Maybe 30fps is a bit optimistic. I will have to think
about that some more. I think the standard web-cam does 320x240 and
640x480 for "hi rez". Maybe 15fps at the lower res will be palletable
(I'll have to experiment).
This information determines your system requirements, and requirements
for the parts of the system (camera, wireless end, what processing e.g.
compression, etc.).
You have to get this data at 30fps from the camera via USB to your
very cheap system, then out again over the same USB controller
to a USB to wireless.
Work out how much data continuous video is and whether you have the
available bandwidth in BITS per second to achieve your requirements
bearing in mind the overheads of USB and Wireless AND ethernet.
What is it with people that think USB is cheap therefore it is
the answer to EVERYTHING!
Because when you don't have a lot of cash to blow you don't have a lot
of choices. In my case I'm trying to stick with what I've got laying
around the house. It doesn't need to be perfect or pretty. Even if I
get this project to simply limp around I will consider a great
accomplishment. How's that for setting the bar high.
It is fine to use what you have if it is capable of doing what you need.
"If all you have is a hammer, everything looks like a nail".
The parallel port is not for printing but for throwing switches. I
supose this could be done through a USB/serial port, but I am not much
of a hard-ware guy, so I don't want to mess with programming a PIC to
translate a serial command to throw a switch.
I would actually say whatever you do you will _probably_ need to write some
software and depending on how the 'throw a switch' translates into hardware
interfacing could be an issue in creating interface hardware. If nothing
else to take some commands from a web form to execute a CGI programme
to then access the parallel port. This turn will need hardware to interface
to your 'switches' whatever they may be.
No 'solutions' I have seen do both video transmission and control up stream
except to control specific pan/tilt/zoom camera protocols. These are usually
specific to each camera.
I would look at telemetry devices and see if you can find a telemetry device
that works one way on a different frequency to send a serial byte to a
telemetry device with parallel o/p. I am sure they exist and relatively
cheaply.
It all depends on what the 'throw a switch' means as in how many, what the
'switch' is controlling in terms of load voltage/currents etc..
It seem I could go several routes. I could go the Linksys route and
find a WAP for about $30. But I don't know about adding USB ports to
it. Or I could go the "broken" laptop route and shove the guts into my
box. But maybe someone from this group could suggest a SBC that will
do what I'm looking for.
Possibly see if you can just simply find an ethernet or wireless webcam
that already exist!
Ethernet cameras ALREADY exist!
Wireless cameras (many NOT using wireless ethernet) ALREADY exist!
Yes but I have a USB web cam.
Which may be part of the problem, I know you have it but it may be easier
cheaper (in money and time) and much quicker to use a better tool.
This data is being sent to some central point (no doubt to connect to
internet for a school project), which will have all the grunt power
required and more capability to run things like apache/php EASILY.
Ah... That's where you're wrong (with all due respect). This project
IS the central point. I can see why you think I'm a student (due to
your preceived short-comings of me)
Hmm this may be central point but I assume the wireless is connecting
to something. If you don't need this accessible via the internet I would
still consider the wireless CCTV to a TV/VCR option. Unless you need to have
the video on another PC for other reasons.
The main thing that made it look like a school project is the method
and requirements. Too often I see
must have x86
must use a cheap web cam
must have wireless
The standard ingredients of a school project.
- but no - I am just a software
developer who was fortunate enough to find a really cool robot platform
at a garage sale for $2, and who is helping out at "robot camp" in
about two months (two of my nephew's are visiting grandma for "robot
week"). We have a couple of old radio shack "Arm-a-trons" and small
"robot" kits for them to build - as well as the obligatory showing of
the animated "Robots" etc. For a 7 and 10 year old (and my 1 month
old) it should be quite fun even if this particular project doesn't get
very far.
I would suggest that getting a bought in solution would be quicker, easier
and lighter to power (smaller batteries).
The quicker solution may get more interest from the children.
You have taken the requirements the WRONG way round and started with
x86 and USB then tried to shoehorn what you want to do into it. In reality
shoehorn a fast desktop into a cigarette box!
Well - yeah. I have some basic needs and a small space and limited
time to learn a bunch of new stuff. By the way, I started by
"shoehorning" an old desktop but there just wasn't enough space - and
then of course the power requirements.....
Thanks again - you didn't really give me the answer I was looking for,
but it did inspire me to keep moving and not wait for the perfect
answer.
With broad requirements one can only give broad answers.
--
Paul Carpenter | paul@xxxxxxxxxxxxxxxxxxxxxxxxxxx
<http://www.pcserviceselectronics.co.uk/> PC Services
<http://www.gnuh8.org.uk/> GNU H8 & mailing list info
<http://www.badweb.org.uk/> For those web sites you hate
.
- References:
- Suggested Starting Point
- From: deja
- Re: Suggested Starting Point
- From: Paul Carpenter
- Re: Suggested Starting Point
- From: deja
- Suggested Starting Point
- Prev by Date: Re: Jump Start PCB design/layout
- Next by Date: Chipcon/TI Dev Board Problems
- Previous by thread: Re: Suggested Starting Point
- Next by thread: Re: Suggested Starting Point
- Index(es):
Relevant Pages
|
