This is merely a historical archive of years 2008-2021, before the migration to mailman3.
A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/OpenBSC@lists.osmocom.org/.
Andrew Back andrew at carrierdetect.comHi Harald, On 04/06/2021 20:15, Harald Welte wrote: > Hi Andrew, > > On Fri, Jun 04, 2021 at 02:38:06PM +0000, Andrew Back wrote: >>> limesdr mini does not have a onboard-clock which is good (precise) >>> enough for gsm requirements. >>> there is no input which can easily be used with an external source like >>> a gpsdo. >>> thus there is no realistic way to reliably do gsm with the -mini, sorry. >> >> External clock can be enabled with LimeSDR Mini, but requires moving >> (soldering) a 0R link and then you always have to use external clock. > > yes,that is possible. I tried that, too, some years ago. The bigger problem, > as far as I recall, are the extremely stringent clock jitter requirements > at this input. There's no on-board PLL anymore, but it was directly the > sampling clock of the ADC, right? Yes, there is no phase detector and ideally the input should be a square wave. There are some more details on the MyriadRF wiki: https://wiki.myriadrf.org/LimeSDR-Mini_v1.2_hardware_description#Clock_Distribution > In any case, at least as I recall, I didn't manage to get a spec > compliant GSM signal out of the LineSDR-mini with that approach using > the clocks I had at my disposal at the time. > > But in any case, that jitter / phase noise is mostly a conformance > related topic and the signal definitely would be good enough to be > properly decoded by commercial UE, particularly in a lab setup with > otherwise near ideal conditions. Sure, for best performance would recommend using the larger LimeSDR-USB or a LimeNET Micro, as Joachim suggested. However, you can use an external reference with LimeSDR Mini and this may provide some performance improvement, but with caveats. Regards, Andrew