Re: [Openbts-discuss] I\Q data handling and buffer overflows as main problem of using OpenBTS with B100

[Kurtis H. <khe...@cs...>]

Interesting. Can you put this on the wiki? It would be good to move people
wanting gprs service away from the b100.


On Thu, Aug 1, 2013 at 2:47 AM, oxccoxcc oxccoxcc <oxc...@ya...>wrote:

> All started from small problem with usual OpenBTS, in debug log,
> periodically, i have seen broken bursts like:(11001101..1010-----...---).
> First part of burst is normal, the second is fully broken.
> This problem is not a critical for OpenBTS main-branch, because just spoil
> a small piece of voice data. But a very critical for OpenBTS-gprs, because
> if it spoils tcp-sessions data, then become a reason of many
> tcp-retransmissions.
> I tried to change daughterboard, it didn`t help. When i tried to use N200
> with original computer, the problem is missed.
> For example:
> ping result with B100=45%-50% packets loss
> ping result with N200=4% packets loss
> N200 and B100 have the same FPGA source code and different ADC\DAC chips
> and output channel(USB and Ethernet).
>
> After that i looked at ERR log and have seen many errors "An internal
> receive buffer has filled ...". USRP B100 send this message when device
> internal buffer was overflowed.
> Because of OpenBTS uses a minnimum possible samplerate(with disabled
> host-based resampling), even minimal I\Q data loss become a reason of
> broken bits.
> Differentiation between N200 and B100, in this case, is that N200 use
> host-based buffer(Linux ring buffer), which is larger than B100
> device-based buffer.
>
> This error often occurs at OpenBTS strating and at variable time when
> OpenBTS is in normal working mode. Error arises even on powerful computers.
> In attach- unmodulated sinusoid. In the time diagram you can see
> drawdowns, when buffer was overflowed.
> Link to attached picture: http://s21.postimg.org/nx5g8bhon/Broken_sin.png
>
> ------------------------------------------------------------------------------
> Get your SQL database under version control now!
> Version control is standard for application code, but databases havent
> caught up. So what steps can you take to put your SQL databases under
> version control? Why should you start doing it? Read more to find out.
> http://pubads.g.doubleclick.net/gampad/clk?id=49501711&iu=/4140/ostg.clktrk
> _______________________________________________
> Openbts-discuss mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/openbts-discuss
>
>