|
From: Thomas M. <tho...@gm...> - 2016-02-10 20:01:41
|
Thanks for the hint. However, I tried with an older version of UHD driver ( UHD_003.010.git-119-g42a3eeb6 and FW Version 8.0 / FPGA Version 14.0) with OpenBTS5 and it still doesn't work... I will try with the 3.9.1 which seems to work for some users right ? Thanks again, Thomas 2016-02-10 8:34 GMT+01:00 Ralph A. Schmid, dk5ras <ra...@sc...>: > Then it is some issue earlier in the chain, not an > authentication/registration problem. I had this when the new UHD with > different group delay came, and it worked again when openbts got patched > respectively. For some reason your setup even does not properly decode the > access bursts. > > > > Ralph. > > > > *From:* Thomas Meghe [mailto:tho...@gm...] > *Sent:* Tuesday, February 9, 2016 6:53 PM > *To:* Ralph A. Schmid, dk5ras > *Cc:* ope...@li... > *Subject:* Re: [Openbts-discuss] Can't register MS on a fresh OpenBTS 5 > install > > > > Apologies for this second email but, I managed to catch access burst on > uplink frequency (cf https://framapic.org/IL7FgT9V35di/YrzHY5d4014s.png) > but there is only one or two packet sent. > > > > 2016-02-09 18:44 GMT+01:00 Thomas Meghe <tho...@gm...>: > > Hi Ralph, > > Thanks for your answer. rxgain was set to 10 (accordingly to the doc > regarding USRP hardware), I tried to set it to 0 which is 10db below. > Nothing is different. > > > > I also tried without the antenna plugged and it still doesn't work. I > didn't manage to find some crappy speakers at home, I will try elsewhere. > However, I can clearly see the downlink of the BTS corresponding to the > ARFCN 80 (which is 951mhz cf > https://framapic.org/zBlgtJB20AfI/EuZAcpYj0duQ.png) but nothing on the > uplink (906mhz) when trying to register MS. > > > > Regards. > > > > 2016-02-09 14:35 GMT+01:00 Ralph A. Schmid, dk5ras <ra...@sc...>: > > What is your rx gain setting? Try using 10 or 20 dB below the preset. Also > try disconnecting your RX antenna when you are very close to the B200 > during the tests. Sometimes the access bursts simply are not decoded > because the signal is too strong. > > > > Can you monitor what the handheld transmits when trying to register? Only > access bursts, or a complete SDDCH exchange? You can get an idea if you put > the phone on EMI sensitive PC speakers or a cheap radio clock, to hear the > annoying GSM noise when it transmits. Access bursts are very short, barely > noticeable clicks, only two or three of them in a row, while an established > SDDCH produces the typical machinegun / drum machine "tack tackatack > tackatack" sound, like when receiving a short message in a commercial > network. > > > > Ralph. > > > > *From:* Thomas Meghe [mailto:tho...@gm...] > *Sent:* Tuesday, February 9, 2016 12:33 PM > *To:* ope...@li... > *Subject:* [Openbts-discuss] Can't register MS on a fresh OpenBTS 5 > install > > > > Hello, > > I'm currently trying to setup a working OpenBTS 5 testing setup (for > testing purposes) on a Ubuntu 14.04 64 bit with a B200. What I've done so > far : > > - Setup OpenBTS 5 following the guide "Getting started with OpenBTS > guide" > - Everything seems to be functionnal (no compile error, no running > error, radio com with my B200 operational) > - Kept MCC and MNC to testing default values (001 01) > - Control.LUR.OpenRegistration set to *. > - Control.LUR.SendTMSIs set to 1 > - sipauthserve running > - noise RSSI is -79 dB wrt full scale > - power downlink is set to -10db > - ARFCN is set to 80 (checked before with OsmocomBB scan that the > associated frequency was not noisy, I can see from my osmophone a nice and > clear peak) > > I think I didn't missed anything for the crash-test but... I tried to > connect on 3 different phones with 3 different simcards (My current one on > my Android phone, an old one and a foreign one). All my 3 phones can see > the network but no one is able to connect to. The connection process is > long and fall short to a "not available" or something else. > > I'm using my B200 with two Vert 400 antennas (it shall be enough to cover > the living room area) > > Thanks for any help and regards, > > Tom > > > > > |