Hello Marcel,
On Wed, 08 Apr 2009 18:19:20 +0200, "Marcel Klein" <marcel(a)koeln.ccc.de> wrote:
>
> Okay now I have access to some more cell phones. But still the same
> issue even with a Nokia 3330 and Nokia 6210.
>
> Sony Ericson k800i (no branding) - doesn't find the network most of the
> time or rather takes very long (when I'm lucky). Only one network in
> list when it manages to find my network. (Just happened once that
> everything was correct).
>
> Nokia 3330 and 6210 - Finds the network quick or just needs to be
> rebooted. Only one network in list.
>
> Siemens S45 - Network not found (yet)
>
> I already tried several networks that really exist but without success.
> ARFCN is free and stable - I checked this with a radio.
I am rather sure this is an issue with the BS-11 clock. It requires a
very accurate and stable clock for GSM (in my tests with just one
phone a BTS clock detuned by about 1 ppm causes the phone to no longer
see both test network and official network). There are most certainly
differences between the phones too in regards to how much inaccuracy
they tolerate.
You can make a test to see if it is a problem with the clock: take a SIM
card from a phone which is registered (location updated has been done)
to the BS-11 network and put it in a phone which did not find the
BS-11 network. Now the phone should be able to see the BS-11 network
when powered on because it immediately listens on the ARFCN of the
BS-11 and does not register (and synchronize its clock) to an
official network first. However it most certainly will no longer
find the official networks.
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
Hello Marcel,
On Wed, 01 Apr 2009 15:52:21 +0200, "Marcel Klein" <marcel(a)koeln.ccc.de> wrote:
>
> Some phones can find the "001 01" network directly but most of the time
> it takes around an hour until some phones start to notice it (or they
> never do). I tried restarting bsc_hack several times, changed the
> ARFCN, MNC and MCC but without success. Even restarting the phone
> doesn't always help.
>
> So far everything seems to work fine, bsc_hack shows no warnings, I see
> lots of RXs when I use the debug option, and Abis-Link is up.
>
> I don't know if this is related but when a phone manages to see my
> network it's not able to see any others. I just see one - my own network.
You mean you can only see the "001 01" network and nothing else ?
Does this happen before or after having registered to the "001 01"
network ?
Something different to try: If finding the "001 01" network takes too
long and changing the ARFCN also does not help (to make sure that
there is no interference) maybe increasing the TRX power changes
the result ? Just try it for a short time (one or two minutes)
and see if that helps. Here at my place (most certainly a rather
"clean" RF environment) I don't have problems to find the "001 01"
network, this works with different phones (old and new ones) and only
takes about 30 seconds doing a "manual search". I can also see all the
other official networks.
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
Hey Harald,
I didn't know your favorite way of getting patches so for now I simply
attached them to this mail. I found some time to continue with channel
management and would like to get feedback if this is running into the right
direction or if you have something else or other ideas.
The reason to not just apply these patches are within the first patch.
Currently we don't serialize the struct gsm_bts so after a restart of the
bsc_hack we would have to force another location updating request...
happy hacking and feedback would be welcomed
z.
Hello guys,
I'm a bit familiar with Linux, but not heavily experienced. Anyway, I
tried to get OpenBNC working but unfortunately with no succes. After
running the commands:
aclocal
autoconf
automake --add-missing
./configure
make
I get after running "make" the following errors:
libbsc.a(db.o)(.text+0x90d): In function `db_sms_mark_sent':
/home/nordin/Downloads/trunk/openbsc/src/db.c:447: undefined reference to `dbi_conn_queryf'
collect2: ld returned 1 exit status
make[1]: *** [bsc_hack] Error 1
make[1]: Leaving directory `/home/nordin/Downloads/trunk/openbsc/src'
make: *** [all-recursive] Error 1
This is the last part of the make output.
Befor this I had the following error running ./configure:
libc6-dev
libdbi-dev
libdbd-sqlite3
Comaplaining that these were missing.
I installed "libdbi-dev" with yum (yum showed me: libdbi-devel.i386, so I thought it's the same)
I than downloaded libdbd-sqlite3 from sourceforge (libdbi-drivers-0.8.3-1 <http://sourceforge.net/project/showfiles.php?group_id=65979&package_id=6377…>) and installed it too.
Finally I downloaded libc6-dev form http://packages.debian.org/search?keywords=libc6-dev, but don't know what to do with it.
Also couldn't find any usefull information on the net, except when using Debian.
I use Linux distro CentOS 4.4 (server edition).
Can you please help me with this?
Thank you very much.
hi,
everything works now, even without external (accurat) clock from telephone network.
the problem was the driver for some reason. with the latest GIT commit it works every time i startup the base or the software. (until now)
now comes the next part. what has to be done in order to terminate calls or make calls between mobiles?
here is the exact way on how to bind (open) mISDN time slots or bchannels: the "dev" field in the sockaddr_mISDN must equal the port number. in case of ony one card installed, it must be set to 0. the "channel" fiels must equal the time slot. because the base station uses time slot 2 and 3 for traffic channels, we must set it to 2 or 3. this is done correctly in misdn.c! is there any problem with getting data from BTS? what is the current state?
best regards,
andreas
-----Ursprüngliche Nachricht-----
Von: openbsc-bounces(a)lists.gnumonks.org [mailto:openbsc-bounces@lists.gnumonks.org] Im Auftrag von Dieter Spaar
Gesendet: Freitag, 3. April 2009 15:08
An: openbsc(a)lists.gnumonks.org
Betreff: Re: WG: BS-11 runs, but no Network on my Mobile
Hello Andreas,
On Fri, 3 Apr 2009 12:27:05 +0200, "Andreas.Eversberg" <Andreas.Eversberg(a)versatel.de> wrote:
>
> so, if this works, i will try to use an hfc-pci single port card
> (cheap) and provide a document on how to link the E1 card to the
> single port card. only two clocks are required, a frame clock and a
> bit clock. with the right settings, both clocks are compatible.
Great idea ! I am definitely interested in the results.
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
i tried different network and country codes (other than '1'). also i tried different ARFCN, like 120 or so.
also note that eary tests showed me the network "01" just once, but at this time i did not get complete provisioning of BS-11 (only 3 channels were configured).
does BS-11 need to react on telephone messages in order to show the network to the phone? or does the phone just receive the BS-11 signalling information in order to display the available network? i don't know if the receiver works.
also i can create BBSIG1 and PA1, but it will not be used by bsc_hack. what do i need to change in order to use the second transponder? maybe this one works.
-----Ursprüngliche Nachricht-----
Von: Dieter Spaar [mailto:spaar@mirider.augusta.de]
Gesendet: Mittwoch, 1. April 2009 15:09
An: Andreas.Eversberg
Cc: openbsc(a)lists.gnumonks.org
Betreff: Re: AW: AW: BS-11 runs, but no Network on my Mobile
Hello Andreas,
On Wed, 1 Apr 2009 12:36:15 +0200, "Andreas.Eversberg" <Andreas.Eversberg(a)versatel.de> wrote:
> the load is correct. i get an error pop-up while starting bsc_hack. this
> is because the LMT seems not to be 100% compatible with the load. with
> an older load i don't get this error pop-up, but it also doesn't work.
> also no alarms show on the LMT after syncing database. i will get the
> exact error message this weekend.
The error message that LMT is not compatible to the firmware version
should be no problem, it just complains about unknown attributes.
Its strange, because you already know that the transmitter
is working, so I wonder why you don't receive something on
the phone. Have you tried a different phone, just to be
sure ? Maybe also using a different ARFCN might help.
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
you can set the clock of BS-11 to internal oscillator. then you
need to set the E1 card into clock slave. (not pcm slave)
modprobe hfcmulti port=0x200 debug=0x40000 # see hfcmulti.h or http://www.linux-call-router.de/download/driver/hfc_multi-dsp-2.0.pdf
and watch the system messages. you can see that the card will get
the clock from the E1 interface and uses it for internal processing.
if you like to get the clock from a different card that is connected
to telephone network, you can connect the clocks, and the driver will
automatically switch to received clock. i will see, if i can provide
instructions on how to connect both clocks. also a small application
must keep layer 1 and layer 2 of the isdn line up all the time.
-----Ursprüngliche Nachricht-----
Von: openbsc-bounces(a)lists.gnumonks.org [mailto:openbsc-bounces@lists.gnumonks.org] Im Auftrag von Dieter Spaar
Gesendet: Samstag, 4. April 2009 20:00
An: openbsc(a)lists.gnumonks.org
Betreff: Re: BS-11 runs, but no Network on my Mobile
Hello,
I did a few more tests to check the BS-11 clock accuracy. I took
a phone which is registered to an official network and tried if
I can see the BS-11 test network doing a manual network search.
As long as the BS-11 oscillator is not locked to the E1 clock,
I can reliable find the test network.
After the BS-11 oscillator has locked to the E1-clock, the test
network is only found every now and then. Here with my BS-11 it
takes about 5 minutes till the oscillator is locked, the attribute
is NM_ATT_BS11_CCLK_ACCURACY, "0" means "high accuracy" (locked),
"1" means "medium accuracy" (free running), "2" most certainly
"low accuracy" (not yet warmed up, I have not verified this
value).
Most certainly the BS-11 oscillator is better than the oscillator
of the E1-card, however I don't know if the calibration is still
OK on all BS-11 units. I have not yet found an easy way to measure
the BS-11 clock (without opening the case), the RF signal is always
modulated and I don't have the equippment to measure the frequency
of the modulated and pulsed RF signal in an accurate way. So this
measurement has to wait.
What you can try is to disconnect the E1 line for about 30 seconds,
this should be enough so that the Abis Link goes down. If you
reconnect, I assume it takes a similar time to lock the oscillator
like it did here. So you have a few minutes to start OpenBSC and
check if you can find the test network.
If this works for most units, a possible workaround is to use the RX
clock for TX on the HFC-E1 card instead of its own oscillator (this
way the BS-11 oscillator determines the E1 clock). I tried that here
already (on Windows), and it seems to work. However I don't know if
mISDN can be configured this way or if it requires a small patch.
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
Hello Andreas,
On Mon, 6 Apr 2009 10:55:14 +0200, "Andreas.Eversberg" <Andreas.Eversberg(a)versatel.de> wrote:
>
> finally it works. i had allot of trouble, because the BTSM was suddenly
> locked. no restart, no object recreation worked. i fixed this by doing a
> software download.
You have LMT, right ? If this happens again, it would be interesting
to know if there is some error message available. LMT has a menu
command (sorry, I don't have the exact name available right now)
which allows to read the errors (I think OpenBSC also receives those
errors when it starts up).
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
finally it works. i had allot of trouble, because the BTSM was suddenly
locked. no restart, no object recreation worked. i fixed this by doing a
software download.
also i updated the mISDN driver, so i don't need the patch anymore. now
it works with the latest GIT commit.
i connected the clock from external isdn line card to the e1 card. but i
think it will also work without line clock.
after call setup from my mobile, i get disconnected, because the number
is not know. this is ok, but i get continuous messages about a time slot
(traffic channel) that cannot be .... (next time i write down the exact
message).
EH09: anyone at easterhegg in hamburg this year? (www.easterhegg.de)
i will bring test server and BS-11.
Hello,
I did a few more tests to check the BS-11 clock accuracy. I took
a phone which is registered to an official network and tried if
I can see the BS-11 test network doing a manual network search.
As long as the BS-11 oscillator is not locked to the E1 clock,
I can reliable find the test network.
After the BS-11 oscillator has locked to the E1-clock, the test
network is only found every now and then. Here with my BS-11 it
takes about 5 minutes till the oscillator is locked, the attribute
is NM_ATT_BS11_CCLK_ACCURACY, "0" means "high accuracy" (locked),
"1" means "medium accuracy" (free running), "2" most certainly
"low accuracy" (not yet warmed up, I have not verified this
value).
Most certainly the BS-11 oscillator is better than the oscillator
of the E1-card, however I don't know if the calibration is still
OK on all BS-11 units. I have not yet found an easy way to measure
the BS-11 clock (without opening the case), the RF signal is always
modulated and I don't have the equippment to measure the frequency
of the modulated and pulsed RF signal in an accurate way. So this
measurement has to wait.
What you can try is to disconnect the E1 line for about 30 seconds,
this should be enough so that the Abis Link goes down. If you
reconnect, I assume it takes a similar time to lock the oscillator
like it did here. So you have a few minutes to start OpenBSC and
check if you can find the test network.
If this works for most units, a possible workaround is to use the RX
clock for TX on the HFC-E1 card instead of its own oscillator (this
way the BS-11 oscillator determines the E1 clock). I tried that here
already (on Windows), and it seems to work. However I don't know if
mISDN can be configured this way or if it requires a small patch.
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
