[hpsdr] Potential New openHPSDR Project

[Graham / KE9H]

Phil:

I agree that the little "motherboard" that plugs a BeagleBoard into 
Atlas should be the
first priority.  Just keep the software layered so we can change 
interfaces later.

As far as a PC versus a remote dedicated processor...
The overwhelming advantage of a PC is an amazingly low cost for an 
integrated processor,
display, keyboard, peripherals, processing resources, etc.

It is just that the more I learn, Windows is a horrible realtime 
environment.  Microsoft
would probably say that it is not realtime, and was not intended to be 
used that way.

So either we figure out how to get Windows to behave the way we want it 
to, or
we allow for a signal processing resource that is separate, and use the 
PC as a
thin client GUI.

For instance, even with the BeagleBoard, it will still be cheaper to use 
a notebook/netbook/
laptop/low-end-desktop as the human interface, rather than come up with 
a custom
set of keyboard/pointer-thing/display/touchscreen/audio system, etc.

Just my thoughts on where we need to head.

--- Graham

==

Phil Harman wrote:
> ***** High Performance Software Defined Radio Discussion List *****
>
> Hi Graham,
>
> All  good points.  I think we may be looking  at two different 
> interfaces - a motherboard that takes a BeagleBoard type of SPC and 
> interfaces  it to the Atlas bus.
>
> The second one is where the board can be remoted, via Ethernet; OzyII 
> will enable this.  These SPCs now come with 100T Ethernet  so we could 
> use that as the means of interconnection. However, in that case 
> perhaps an small  PC would be a better option?
>
> I understand that the free TI Code Composer V4.1.2 IDE will support 
> the DSP on these boards but, as you rightly point out, we need to be 
> 100% sure of this before going too much farther.
>
> Please let me know when you are free again since I have plenty more 
> "weekend" projects for you :)
>
> 73's Phil...VK6APH
>
>
> ----- Original Message ----- From: "Graham / KE9H" <KE9H at austin.rr.com>
> To: <phil at pharman.org>
> Cc: <hpsdr at openhpsdr.org>
> Sent: Sunday, July 11, 2010 4:33 AM
> Subject: Re: [hpsdr] Potential New openHPSDR Project
>
>
>> Phil:
>>
>> I am not sure I care which hardware board you select, provided that the
>> processor
>> family has a long life ahead of it, for the reasons many stated. More
>> important that there
>> be a free or low cost tool chain for programming/development. Any card
>> that requires
>> purchasing a USD $1000 or $ 2000 compiler and IDE from the manufacturer
>> is a bad idea.
>>
>> Although having an ethernet interface would be a positive.
>>
>> This seems to be a board that would pull the data straight off the atlas
>> bus.
>>
>> I would like to request that the software be architected properly, so
>> that the processor
>> card could accept data via an interface layer, directly from the bus, as
>> well as,
>> by just substituting a different layer, accept the data via Ethernet
>> cable, or perhaps a serial
>> light fiber driver layer.
>>
>> I like the idea of a dedicated processor card, but I would like the
>> option of
>> having it local on the Atlas bus, or remoted several hundred feet. Does
>> this require
>> a different host/interface board for each application, and still the
>> same computer board?
>>
>> I still have not finished project managing the last "weekend" project
>> you got me
>> into in 2007, (also know as Alex), so I will not be available to join
>> this until after Alex
>> finally gets shipping from the factory.
>>
>> :-)
>>
>> --- Graham / KE9H
>>
>> ==
>>
>>
>>
>> Phil Harman wrote:
>>> ***** High Performance Software Defined Radio Discussion List *****
>>>
>>> All,
>>>
>>> There seems to be some interest in building a  Microprocessor/DSP board
>>> that will plug into the Atlas bus and enable an openHPSDR system to
>>> operate without the need to have a PC connected.
>>>
>>> Whilst one option would be to design and manufacture our own board from
>>> scratch, which would be an interesting and challenging project, 
>>> another is
>>> to build on existing work by others.
>>>
>>> The Beagle board  (http://beagleboard.org/) looks like a suitable
>>> candidate since it has both an OMAP3530 processor and TMS320C64x+™ DSP.
>>>
>>> It also appears that some SDR support is already available in that 
>>> dttSP
>>> has been ported and support for the SoftRock exists.
>>>
>>> In terms of getting data into the DSP from the Atlas bus there is an
>>> McBSP3 interface, which I understand can be configured as I2S, that
>>> connects directly to the DSP.
>>>
>>> The suggestion is that we design a simple, low cost, mother board that
>>> will take a Beagle board and enable it to be interfaced to the Atlas 
>>> bus.
>>>
>>> This would be along the same lines as what the USRP folks are doing 
>>> here:
>>>
>>> http://members.cox.net/ebrombaugh1/embedded/beagle/beagle_fpga.html
>>>
>>> We can also build on ideas from the BeagleBrick announced at Dayton 
>>> this
>>> year:
>>>
>>> http://beagleboard.org/project/BeagleBrick/
>>>
>>> Kevin, M0KHZ, has available to layout a suitable PCB and I'd be 
>>> happy to
>>> develop any FPGA code that may be required.
>>>
>>> Is there any interest in such a project and anyone willing to be the
>>> project leader please?
>>>
>>> 73's Phil....VK6APH
>>>

 1278856095.0