Reply by Didi August 1, 20072007-08-01
> Therefore the hardware vendors force you to treat WiFi as a > black box unless you're an OS vendor or huge-scale OEM.
Thanks, that may well be it. Or it is just the excuse they use for keeping the data secret - or, more likely, both :-). I managed to locate some data, many of the adaptors are compatible to an old Intersil device, which has been sold to some party member and is now secret - but data can still be found on it, I believe. Now I have "only" to find out if that NXP part I have data for is compatible - although via SPI - or find some alternative etc. I'll manage it somehow. All, please keep this in mind and keep the group (or just me) posted if you stumble across some relevant data which seems to be not widely available. Thanks, Dimiter ------------------------------------------------------ Dimiter Popoff Transgalactic Instruments http://www.tgi-sci.com ------------------------------------------------------ On Jul 31, 6:49 pm, larwe <zwsdot...@gmail.com> wrote:
> On Jul 31, 5:40 am, Hans-Bernhard Br=F6ker <HBBroe...@t-online.de> > wrote: > > > Summing it up, WLAN has a "no simple implementations possible" rule > > built right into its design. > > I'm not sure that is entirely accurate characterization. I think > you'll find (certainly the hardware mfrs claim this) that the reason > low-level chipset details are proprietary is because revealing them > would allow you to write software that voids the module's type > approval. (Power and frequency allocation are restricted by geographic > area). Therefore the hardware vendors force you to treat WiFi as a > black box unless you're an OS vendor or huge-scale OEM. > > Bluetooth, by contrast, is the same worldwide (AFAIK).
Reply by larwe July 31, 20072007-07-31
On Jul 31, 5:40 am, Hans-Bernhard Br=F6ker <HBBroe...@t-online.de>
wrote:

> Summing it up, WLAN has a "no simple implementations possible" rule > built right into its design.
I'm not sure that is entirely accurate characterization. I think you'll find (certainly the hardware mfrs claim this) that the reason low-level chipset details are proprietary is because revealing them would allow you to write software that voids the module's type approval. (Power and frequency allocation are restricted by geographic area). Therefore the hardware vendors force you to treat WiFi as a black box unless you're an OS vendor or huge-scale OEM. Bluetooth, by contrast, is the same worldwide (AFAIK).
Reply by July 31, 20072007-07-31
Didi wrote:

> Quite a different story with wifi - to an extent I begin to think > there is something I don't get, some general standard I have not read > or whatever.
I suspect you're looking at the same problem that causes support for WLAN hardware in open-source OSes to be so tricky. The general rule you're missing is that there are some WLAN regulations regarding use of frequency bands and encryption which, according to the general consensus, can only be met by having large parts of the processing done not by firmware, but by run-time selectable drivers. The higher-level drivers acquire the location of the machine, and tell the DSPs which frequency bands are off-limits. Summing it up, WLAN has a "no simple implementations possible" rule built right into its design.
Reply by Didi July 29, 20072007-07-29
The time has come when I have to choose which Bluetooth
and which wifi chip/module to use.
 After may be a whole days browsing/searching, I seem to come
to the following conclusions:
 Bluetooth is available from more than one source as a single package
solution, some of them are documented and really ready to use.
I tend to stop at NSCs part, LMX9830 - stocked and inexpensive
at Digikey, plus that NSC are among the "good ones" in my book.

 Quite a different story with wifi - to an extent I begin to think
there is
something I don't get, some general standard I have not read or
whatever.
 Nothing - 0 - one can buy and use as a peripheral to make conecting
to wifi as much work as say writing a new Ethernet driver plus perhaps
ARP or something.
 Most seem to use a Marvell chip - in another thread all posters
agreed
it is a waste of time trying to talk to them (I guess they will talk
only to party members). So this part is secret.
 A while ago Guy Macon posted a manual for an NXP part which does
contain much more information than anything else, but the part has
some ROM with firmware inside it and while the mechanism to talk
to the internal CPU is documented, the syntax - command set etc. - is
simply not present, so this info is probably secret as well. (Guy,
please let me know if you manage to squeeze some more data at
some point).

 Hopefully there are people around who can prove me wrong so it turns
out wifi is not reserved for party members only.

 Are all those USB/PCC/SD card etc. wifi adaptors documented for MS
and
similar  only? Sounds incredible. Anyone having been able to talk
to such a device directly?

Dimiter

------------------------------------------------------
Dimiter Popoff               Transgalactic Instruments

http://www.tgi-sci.com
------------------------------------------------------