Re: [oc] Re: Need a little direction

[Richard Herveille <richard@asics.ws>]

>
>We want to create the following platform and we would appreciate all help.
>We would even try and support the efforts in reasonable financial ways.
>
>1. BOM should not cross USD $50.
>2. Should have wireless (Bluetooth, short range)
>3. Should have 32MB Flash
>4. Should have 256 MB SDRAM
>5. Should have two PC Card slots.
>6. CPU (could be a SOC)
>
>After doing considerable search, there are some of the options
>1. taking the SA1110 reference design and depopoulate the board extensively.
>2. design from scratch
>
>Now biggest thing to be descided is what SOC/CPU to use. The options are
>(you can suggest more)
>1. StrongArm from Intel
>2. DragonBall MX1 from Motorola (would be mass produced in March 2002),
>having integrated support for bluetooth and cost less than $25, they claim.
>3. AMD Elan processors (they are x86, and they consume as i am told too much
>power, but they are cheaper)

Depends also on the computation power you need, take a look at these too:
1) Motorola Coldfire (e.g. MCF5206e)
2) Intel XScale (next generation StrongArm)

Both are available at the moment. I know for sure that the MCF5206e should 
be available below USD10,=.

>The massive amount of SDRAM (I guess they are faster than flash and lot
>cheaper too) , is for big enough RAM filesystem, which can then download
>content from masterstation, like MP3 and MPEG video etc to be streamed on
>wireless (BT), this content would get lost if power goes off, which we
>believe is not the most often case.(the unit would be externally powered)
>and would have battery for maybe 1 hour operation.

Nasty about SDRAMs is the huge amount of power they consume during refresh. 
You could try to look at Micron's (www.micron.com) BatRAMs which are 
special SDRAMs for battery operated devices.

Most SoC processor do not feature a SDRAM interface, but that could be 
easily implemented in an FPGA. You need some glue-logic anyway. Also 
instead of using PC-CARDS look at the Compact-Flash devices. They are much 
smaller, lighter and use less power. But above all, everything memory size 
available for PC-CARD is also available for CompactFlash. IBM's MicroDrive 
is a CompactFlash type-II true harddisk (up to 1GigaByte !!!!!). Also very 
nice about the CompactFlash is that is can be controlled using a simple IDE 
interface, the OCIDEC-1 core available from OpenCores is ideal for this 
purpose.

For an FPGA I would recommend Altera's APEX, or Xilinx's Spartan-II devices.

To be honest sounds almost like a board I developed :-)).

Richard

>What we need is
>1. freelancers to work on the designboard or any volunteers
>2. pointers to information and any feedback to help us achieve what we set
>out to achieve. Any criticism or any sort of feedback is appreciated.
>3. Component manufacturers/contract manufacturers who can give help us
>achieve that price.
>4. any kind of support.
>
>We are based in India (Bangalore), and if anyone happens to be in Bangalore
>and want to drop by our place you are more than welcomed to do that. Any
>enquires please send it to me
>
>Looking forward for more pointers , feedback, answers in search for the
>perfect solution.
>
>Best Regards
>Devesh Kothari
>Director
>ProtoSys Technologies Pvt Ltd
>www.protosys.com
>dkothari@protosys.com
>22/2 Mission Rd, Samson Complex,
>Bangalore, 27, INDIA
>
> >
> >  > [..] I am complete novice in terms of
> >  > hardware (please dont suggest to just drop it then).
> > I will not. I do not know what you are into exactly. Nothing is
>impossible.
> >
> >  > Then I would start the porting on one of Linux variant on the box.
> > You may know that Linux has already been ported to the Strong Arm
> > architecture and Debian for example is providing a complete
> > distribution.
> >
> >  > [..] I have
> >  > all the BOM/Gerber files and schematic diagram from intel, but not sure
>how
> >  > to proceed.
> > I guess that you want to use the gEDA tools and other free tools and
> > that is why you mailed me.
> >
> > If so, basically you will have to reenter all the schematics from Intel
> > into gschem. That mean that you will also have to create a new symbol
> > for almost all the components, i.e. SA1110, bluetooth module ...
> > Then you will have to generate a netlist from the schematic - with
> > gnetlist.
> > And finally, you will have to route the board. I am not sure that the pcb
> > software from Harry Eaton would be complete enough for this task. In fact,
> > SA1110 is quite a complex component as it is in a BGA - Ball Grid Array -
> > package. At the working frequency you also have to take into account
> > the way the board is tracked.
> >
> > Of course, that was the scheme for the software manageable part. You will
> > have to place an order for this board to a pcb maker.
> > As the package for SA1110 is a BGA you will not be able to solder it by
> > yourself and will require external help for that.
> >
> >  > It would be quite useful if someone could help us finnish this
> >  > project. If you can provide us with some direction please.
> > I would be pleased to help you as much as I can. Let me know how far
> > in the specification of the device you are, what kind of direction
> > you need.
> >
> >  > Maybe I am just wasting your time. In that case please accept my
>apologies.
> >  > Sorry for the trouble
> > I am not sure to have answered all your questions. If my explanations
> > were too basic, excuse me.
> >
> >
> > Patrick BERNAUD
> >
> >
>--
>To unsubscribe from cores mailing list please visit 
>http://www.opencores.org/mailinglists.shtml

--
To unsubscribe from cores mailing list please visit http://www.opencores.org/mailinglists.shtml