Hallo Kai,
On Tue, 23 Jun 2009 14:19:20 +0200, "=?ISO-8859-15?Q?Kai_M=FCnz?=" <squ(a)tent.at> wrote:
>
> There is an even lighter setup possible.
>
> I'm using an old IBM Thinkpad R50p with a Thinkpad DockII (Ebay, 30Eur)
>
> http://www.bildschirmtext.org/thinkpad-e1.jpg
>
> Just connect the BTS and play %)
Nice, and it even has DECT connnectivity (looks like a COM-ON-AIR card
in the background :-)
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
Hello Andreas,
On Tue, 23 Jun 2009 13:30:17 +0200, "Andreas.Eversberg" <Andreas.Eversberg(a)versatel.de> wrote:
>
> picture!
Even more cool :-). I expected a bigger car, maybe a van, to but all
the stuff inside. Really nice GSM war-driving setup ;-)
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
Hello Andreas,
On Tue, 23 Jun 2009 10:12:38 +0200, "Andreas.Eversberg" <Andreas.Eversberg(a)versatel.de> wrote:
>
> how can i advice the phone to use the "full" power level according to
> the power class of the bs11? what parameter and what value is required?
Currently the BS power is set in rsl_rx_chan_rqd() (file abis_rsl.c),
its the "lchan->ms_power" assignment. Currently the value 0x0F (13 dBm)
is used (I am not sure if this is still true for the latest version of
OpenBSC). If you set this value to 0, the MS will use the maximum
power (usually only 5 is supported for GSM900, 2 Watt). I did only
a few tests, but I could see that my phone uses the modified value.
You can of course also look at the measurement report and check the
used values.
> currently we have "kieler woche", an event with up to 3.000.000 visitors
> over one week.
>
> for the "mobile base station" i use a standard PC with E1/S0 card, a
> BS11, a laptop, an ISDN phone, and an UPS with 4 * 200Ah batteries.
Cool, do you have some pictures of the setup ? :-)
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
Hello Andreas,
On Mon, 22 Jun 2009 19:40:17 +0200, "Andreas.Eversberg" <Andreas.Eversberg(a)versatel.de> wrote:
>
> after location update fails. "DB: Failed to find the Subscriber...", get
> an "ERROR INDICATION" with cause 1. it seems that the mobile just stops
> sending on the channel. the channel ressource hold by location update
> process is not freed. the last messages show that.
>
> i will look on this the next days and report if i found something. if
> you have any idea, please tell me. the way to test any change/fix is
> quite complicated. the bug only occurrs when many phones are available
> and when they move from a different network to my network (built in a
> car, moving arround.)
Just a wild guess: Could it be a reception problem ? If you have set
the BS-11 to a high power level, many phones will see it. Currently the
BS-11 advises the phones to use a very low power level when activating
a channel. So the BS-11 might receive the phone with a low signal
strength and the receiption is just too bad (although the phones will
receive the BS-11 quite well). The measurement reports should indicate
if this is the problem.
Out of interest, do you also have a PC with an E1 card in the car or
do you use a laptop with an USB (or PCMIA) to E1 interface ?
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
hi,
i got a leak of channel ressource. (currently i cannot look at the
source code.)
ignore the first 3 lines, they are just part of testing/debugging code i
use.
after location update fails. "DB: Failed to find the Subscriber...", get
an "ERROR INDICATION" with cause 1. it seems that the mobile just stops
sending on the channel. the channel ressource hold by location update
process is not freed. the last messages show that.
i will look on this the next days and report if i found something. if
you have any idea, please tell me. the way to test any change/fix is
quite complicated. the bug only occurrs when many phones are available
and when they move from a different network to my network (built in a
car, moving arround.)
andreas
<8000> chan_alloc.c:164 looking for free signalling subchannel on CCCH
<8000> chan_alloc.c:168 requesting SDCCH* channel
<8000> chan_alloc.c:135 free lchan (1) found on trx 0 ts 1
<0010> abis_rsl.c:894 Activating ARFCN(123) TS(1) SS(1) lctype SDCCH
chan_nr=0x49 r=LOCATION_UPDATE ra=0x15
<0010> abis_rsl.c:744 channel=(bts=0,trx=0,ts=1) chan_nr=0x49 CHANNEL
ACTIVATE ACK
<0001> abis_rsl.c:988 channel=(bts=0,trx=0,ts=1) chan_nr=0x49 ESTABLISH
INDICATION
<0004> gsm_04_08.c:1528 LOCATION UPDATING REQUEST
<0004> gsm_04_08.c:1200 LUPDREQ: mi_type=0x04 MI(1293316309) type=NORMAL
<0002> gsm_04_08.c:347 lchan (bts=0,trx=0,ts=1,ch=1) increases usage to:
1
<0004> gsm_04_08.c:1226
<0002> gsm_04_08.c:990 (bts 0 trx 0 ts 1 pd 05) Sending 0x18 to MS.
DB: Failed to find the Subscriber. '1' '1293316309'
<0002> gsm_04_08.c:990 (bts 0 trx 0 ts 1 pd 05) Sending 0x18 to MS.
<0008> gsm_04_08.c:1250 <- Can't find any subscriber for this ID
<0001> abis_rsl.c:988 channel=(bts=0,trx=0,ts=1) chan_nr=0x49 DATA
INDICATION
<0008> gsm_04_08.c:1634 CLASSMARK CHANGE CM2(len=3) CM3(len=4)
<0001> abis_rsl.c:988 channel=(bts=0,trx=0,ts=1) chan_nr=0x49 ERROR
INDICATION cause=0x01
<0001> chan_alloc.c:256 Recycling the channel with: 0 (0)
<0010> abis_rsl.c:530 Channel Release CMD channel=(bts=0,trx=0,ts=1)
chan_nr=0x41
<0010> abis_rsl.c:744 channel=(bts=0,trx=0,ts=1) chan_nr=0x41 RF CHANNEL
RELEASE ACK
<8000> chan_alloc.c:203 freeing logical channel on trx 0, ts 1
<0010> abis_rsl.c:744 channel=(bts=0,trx=0,ts=1) chan_nr=0x49 CONNECTION
FAIL: CAUSE: 18 01 49 Cause 0x18 IGNORING, lchan in use! (1 times)
<0010> abis_rsl.c:744 channel=(bts=0,trx=0,ts=1) chan_nr=0x49 CONNECTION
FAIL: CAUSE: 18 01 49 Cause 0x18 IGNORING, lchan in use! (1 times)
<0010> abis_rsl.c:744 channel=(bts=0,trx=0,ts=1) chan_nr=0x49 CONNECTION
FAIL: CAUSE: 18 01 49 Cause 0x18 IGNORING, lchan in use! (1 times)
<0010> abis_rsl.c:744 channel=(bts=0,trx=0,ts=1) chan_nr=0x49 CONNECTION
FAIL: CAUSE: 18 01 49 Cause 0x18 IGNORING, lchan in use! (1 times)
Hello guys,
I was studying the sourcecode of OpenBSC to get a good understanding of
how things work.
But due to my lack of experience in Linux programming I have some
difficulties to understand the source.
I understand that the way bsc_hack works is based on event-queue
concept, am I right?
Anyway, in select.c I don't understand what's going on when it comes to
registering and unregistering fd's in combination with linuxlist.c.
Can someone please give me some links where such concepts are being
explained, so I can study it?
Thanks in advance.
Hello Harald,
On Sun, 21 Jun 2009 01:05:27 +0200, "Harald Welte" <laforge(a)gnumonks.org> wrote:
>
> mh. I see. Seems like there is some more research required here. With
> the recent "L1 Info" IE decoding that I committed, we should always see
> the actual RF power in dBm used by the MS during reception of measurement
> results.
The MS power can easily be observed with the Network Monitor of the
Nokia phones, the updated frequency of the display is fast enough.
I can for example see that the MS POWER of the ACTIVATE CHANNEL
command is set (as expected of course).
> No, this is actually the inverse test: MAke sure the power does _not_ change
> if TCH and BCCH are on the same TRX and we send different BS POWER values
> for the TCH CHANNEL ACTIVATE. Right now we send a value of 15 (!).
The BS power of the BCCH does not change. From my understanding, this
is the expected behaviour because the BCCH frequency is the reference
of the cell for all such things as cell re-selection or handover. I
think (without being 100% sure) that this even affects all other
timeslots of the BCCH TRX because if a phone measures the signal
strength of a neighbor cell, it does not neccessarily measure on TS0
(not sure how it is acually done, RSSI average over a certain amount
of time ?)
I cannot confirm with a measurement that the TCH power is not changed
because I cannot measure it if TCH and BCCH use the same frequency.
BTW, what does the MS measure and report in the measurment report if
there is an acctive connection, is it the strength of the TCH or the
BCCH ?
> CCCH is the channel that contains the BCCH. So I'm actually asking for
> what you "don't think", i.e. an attempt to alter the BS POWRE on the TRX
> that carries the BCCH.
As I said, I don't see a change. But as long as it is not 100% sure what
I or the phone measures (BCCH or TCH) there is still a chance that the
TCH is adjusted, although I don't expect it (not that it isn't possible,
I have read that on a TRX which does not carry the BCCH, a different
power level on every timeslot is allowed).
One other observation: I tried to use a different ARFCN for the TCH
on the BS-11 using only one TRX. This does not seem to work, although
I can see with LMT the ARFCN set for all the TCH timeslots and can
also see that the phone switches to the other ARFCN, there is no RF
activity of the BS on this ARFCN. All the commands to activate the
channel are acknowledged and use the other ARFCN, no errors. The
only strange thing is the ARFCN list of the TRX, I have added the
other ARFCN but LMT displays "0" for all additional ARFCNs in the
list, only the first has a different value.
Maybe I am doing something wrong or this is just the expected
behaviour, a TRX which carries the BCCH cannot use a different
ARFCN for the TCH (but this would also mean that no hopping is
allowed).
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
Hello Harald,
On Sat, 20 Jun 2009 10:02:17 +0200, "Harald Welte" <laforge(a)gnumonks.org> wrote:
>
> Do you get an error message (SET ATTRIBUTE NACK) if you try to set it?
I have not looked at the response, but LMT says "unrecognized value"
or something like that if I query the TRX attributes. The BS-11 seems
to use "6" if something larger than 6 is sent. The nanoBTS 1800 however
does not start if an invalid value for NM_ATT_RF_MAXPOWR_R is sent,
the green LED is just blinking.
> According to the spec, 6 steps of 2dB is the minimum a BTS has to support.
Yes, and looking at the test report of the BS-11 seems to indicate that
there are 15 steps possible for dynamic adjustment (probably used only
if BTS power control is enabled).
> sure. Also, the BTS is 12 years old...
I think the power of the BS-11 is still rather accurate, someone
else (you know him) with much better measurment equippment has
confired a while ago that the maximum power of the BS-11 is very
accurate. Anyway, 3 dB more or less is not really a problem for
our purpose.
> MS power control (the dynamic adjustment of MS power) should be used even now,
> otherwise I would not understand my observation of the phone bursts becoming
> much lesss loud in the speakers after the initial few very loud bursts.
I can only report the results of a short test. According to LMT, the
MS power control is currently disabled when bsc_hack is used (at least
the version I use). If I enable it and additionally enable the whole
power control of the BS-11, I can see that the MS power is changed.
I guess the same is true for the BS power.
> What would be more interesting to me than dynamic BS power control is:
> How do the 'BS POWER' IE's in the ACTIVATE CHANNEL and 'BS POWER CONTROL'
> messages affect the BS transmit power.
>
> Some things to confirm:
>
> 1) whatever we use as BS POWER value in ACTIVATE CHANNEL on a TCH/SDCCH8
> on the C0 does not make any changes to the acutal TX power
Difficult to measure for me with the current setup (ARFCN of the
traffic channel is the same as the BCCH channel). To find out if the
BS power of the traffic channel is modified, I have to switch to a
different ARFCN for traffic.
> 2) if we activate a channel on the second TRX, do we see the BTS power
> adjusted according to BS POWER in ACTIVATE CHANNEL ?
Does bsc_hack already support the second TRX ?
> 3) if we use a BS POWER CONTROL message on the CCCH on C0 of an otherwise idle
> BTS, do we see a power change on the TRX ?
I can only measure it if I switch the ARFCN to a different channel than
used by the BCCH. I don't think the BTS will modify the BCCH power.
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
On Fri, 19 Jun 2009 10:55:16 CEST, "Dieter Spaar" <spaar(a)mirider.augusta.de> wrote:
>
> I will see if I find some time for measurements with the BS-11.
And here are the measurment results for the BS-11 for each of the
four power classes:
BS-11, ARFCN 123
BTS Power: 0.03 Watt
NM_ATT_RF_MAXPOWR_R RF output
0 12 dBm
1 9.7 dBm
2 7.8 dBm
4 4.7 dBm
6 0.9 dBm
BTS Power: 0.08 Watt
NM_ATT_RF_MAXPOWR_R RF output
0 17 dBm
1 15 dBm
2 13 dBm
4 8.7 dBm
6 5.5 dBm
BTS Power: 0.25 Watt
NM_ATT_RF_MAXPOWR_R RF output
0 22 dBm
1 20 dBm
2 18 dBm
4 14 dBm
6 9.9 dBm
BTS Power: 2 Watt
NM_ATT_RF_MAXPOWR_R RF output
0 32 dBm
1 30 dBm
2 27 dBm
4 23 dBm
6 19 dBm
Values larger than 6 for NM_ATT_RF_MAXPOWR_R are not supported.
There is most certainly an error in the range of 2 to 3 dB coming
from the low-quality cable and some adaptor connectors to connect
the BTS to the measurement equipment.
I did also play with the power control of the BS-11. As far as I
am aware the BTS power control is not enbabled per default in
bsc_hack, the measurement results from the BTS confirm that. If
I enable the BTS power control I can at least see that the BTS
changes the MS power. I have not verified on the phone if the
power is really changed, but the RX level of the BTS seem to
indicate that it works.
There are a lot of parameters to play with so if anyone is
interested, just reduce NM_ATT_BS11_RADIO_MEAS_GRAN (lets
say to 2) so that you see frequent measurement results and
watch what is going on.
Best regards,
Dieter
--
Dieter Spaar, Germany spaar(a)mirider.augusta.de
