OLPC hardware: what if there was an SDR modem / chipset?

[Luke Kenneth Casson Leighton]

On Tue, Jan 26, 2010 at 3:22 AM, John Watlington <wad at laptop.org> wrote:
>
> On Jan 26, 2010, at 2:07 AM, Luke Kenneth Casson Leighton wrote:
>
>> On Mon, Jan 25, 2010 at 6:05 PM, Luke Kenneth Casson Leighton
>> <lkcl at lkcl.net> wrote:
>>
>>>  the key questions to be asking are: in light of the massive volumes
>>> involved with OLPC XOs, is SDR worth pursuing, given all the
>>> development costs, but given all the benefits especially where SDR can
>>> be re-programmed to do "whatever" in bands where licensing and
>>> regulatory approval is _not_ required, on a per-country basis?
>>
>>  bearing in mind that the particular R.F. front-end transceiver chips
>> i'm looking at, one can do 600mhz to 6ghz and the other can do 100mhz
>> to 6ghz.  there's another one which can do 80mhz to around 5ghz or so
>> i believe, but it's "receive only" but has far superior
>> signal-to-noise ratio and is accompanied by a far superior A2D-USB2
>> converter.
>
> A USB2 interface is used to talk to the A/D ?

( on the mirics receiver chip, yes, which is no good cos it only
does... receive!  i only mentioned it for completeness.  i did say
"but" ).  i don't know the details of the other two, yet, but they're
likely to be memory-addressed.

> I just eliminated all the internal USB buses in the XO...
> I'm not against SDR, but this chip doesn't sound
> compatible with our power/cost budget.

 a)  there are three separate and distinct chipsets that i've found.
 b)  we don't _know_ yet what the exact power budget is
 c) we don't _know_ yet what the exact cost budget is.

 so it's early days.  the main things i want to establish are "what
are the criteria for acceptance".  you've said "cost and power".  so,
i now have enough to go ask the right questions.

> More importantly, this sounds like it is years
> of software development and testing from being
> a real product.

 mmm... no, i don't believe so.   lyrtech already have the necessary
SDR expertise, for example, and have released white papers describing
the benefits of turning GSM handset SDR from matlab simulations into
"real" hardware, actually doing the development on their SDR platform
to prove it (translation: they actually _have_ GSM SDR software).  i
also know a company with an SDR 3G turnkey solution.  both companies
have WIMAX in SDR.  also, the research companies who have been
actually doing the RF chips also have to "prove" that the chips
actually work, both to themselves and to potential clients, and that
means using "real" 3G and other "real" software, not just doing a few
RF test plots.

conclusion: software is already available.  therefore, software
development is not years, it's months.

>   OLPC doesn't have the resources
> to fund this development.

 i didn't say "fund".  "fund" was _specifically_ left off the list.
check again what i wrote.  i said "off-the-shelf", with the option to
re-program WITHOUT requiring NDAs.

(and i also specifically mentioned the gnu-radio project resources as
being one option.  there are a _lot_ of people who would be
_extremely_ happy to not have to spend $750 on a USRP in order to do
free software development gnu-radio.)

> Re certifications, on both XO-1 and XO-1.5 we
> use slightly non-standard modules (on XO-1 to
> support 802.11s, on XO-1.5 to reduce power
> consumption by 50%) and have paid the full
> certification costs.   They run about 100K$ for
> starters, with each additional country adding
> around 5K$.

 ok.   interesting to know.  ridiculously high, but important to know.
 and, given that they _are_ non-standard, such that the full
certification costs would have to be paid, that means that you're "in
the game" already for paying such up-front certification costs if
there was a ready-made off-the-shelf SDR modem kicking around.

l.