[Solved] YateBTS and bladeRF running problem

[andrew77]

Hello guys,
I have a problem when running yate with bladeRF:
I have USB2.0
the installation of yate and yate BTS all gone OK via SVN

this is the output when i run:
 yate -Df -vvvv -l /var/log/yate.log

20150411113607.759068 <javascript:INFO> Parsed 'nib' script: /usr/local/share/yate/scripts/nib.js
20150411113607.763468 <nib:CONF> Please configure country code. See subscribers.conf or use the NIB web interface
20150411113607.765483 <NOTE> Failed to open config file '/usr/local/etc/yate/tmsidata.conf', using defaults (2: No such file or directory)
20150411113607.766009 <nib:INFO> Finished reading saved registered subscribers. Found 0 registered_subscribers.
20150411113607.766748 <nib:INFO> Checked that only NIB is enabled.
Yate engine is initialized and starting up on and-HP-Compaq-2510p-Notebook-PC
20150411113607.767043 <INFO> Creating first message dispatching thread
20150411113608.279609 <cpuload:NOTE> Updating CPU core number from 1 to 2
20150411113608.486824 <mbts:MILD> TRXManager.cpp:283:sendCommandPacket: TRX link timeout on attempt 1
20150411113609.487476 <mbts:MILD> TRXManager.cpp:283:sendCommandPacket: TRX link timeout on attempt 2
20150411113610.488483 <mbts:MILD> TRXManager.cpp:283:sendCommandPacket: TRX link timeout on attempt 3
20150411113611.489554 <mbts:MILD> TRXManager.cpp:283:sendCommandPacket: TRX link timeout on attempt 4
20150411113612.490525 <mbts:MILD> TRXManager.cpp:283:sendCommandPacket: TRX link timeout on attempt 5
20150411113612.490618 <mbts:NOTE> TRXManager.cpp:293:sendCommandPacket: lost control link to transceiver
ALERT 3075032832 13:36:12.4 TRXManager.cpp:432:powerOff: POWEROFF failed with status -1
20150411113612.491006 <mbts:WARN> TRXManager.cpp:432:powerOff: POWEROFF failed with status -1
20150411113612.595984 <mbts:NOTE> OpenBTS.cpp:145:startTransceiver: starting transceiver ./transceiver w/ 1 ARFCNs and Args:
EMERG 3074321216 13:36:12.5 OpenBTS.cpp:151:startTransceiver: cannot find ./transceiver
20150411113612.596437 <mbts:GOON> OpenBTS.cpp:151:startTransceiver: cannot find ./transceiver
EMERG 3074321216 13:36:12.5 OpenBTS.cpp:159:startTransceiver: Transceiver quit with status 256. Exiting.
20150411113612.597726 <mbts:GOON> OpenBTS.cpp:159:startTransceiver: Transceiver quit with status 256. Exiting

ybts.conf doesn't show transceiver-bladerf:

Path: string: Path to the transceiver relative to where the MBTS is started.
; Should be one of: ./transceiver-rad1 ./transceiver-usrp1 ./transceiver-uhd
; Defaults to ./transceiver.
;Path=./transceiver

but config.log produced by Yatebts's ./configure seems ok:

checking for bladeRF support using pkg-config
configure:4448: result: 1.2.1-2015.02-1-ppatrusty
configure:4483: checking libusb-1.0/libusb.h usability
configure:4483: gcc -c  -O2 -DLITTLE_ENDIAN  conftest.c >&5
configure:4483: $? = 0
configure:4483: result: yes
configure:4483: checking libusb-1.0/libusb.h presence
configure:4483: gcc -E  conftest.c
configure:4483: $? = 0
configure:4483: result: yes
configure:4483: checking for libusb-1.0/libusb.h
configure:4483: result: yes
configure:4749: creating ./config.status

I don't know how to solve this issue

[Vitia]

Hello,

I was there too, but the "NIB" will set a new entries in that ybts.conf file:

[transceiver]
;Path to the transceiver relative to where MBTS is started.
;Should be one of: ./transceiver-bladerf ./transceiver-rad1 ./transceiver-usrp1 ./transceiver-uhd.
;Defaults to ./transceiver
Path=./transceiver-bladerf

Pls, refer to http://wiki.yatebts.com/index.php/Javascript_NIB

I would like to exchange with you regarding "noise" on the bladerf afterwad... I have a half running yatebts/bladerf x115 (sms ok, phone registration ok, but unstable...)

Good luck and please report your status...

Vitia

[andrew77]

OK I will try to configure NIB

Should I install Nib? It seems already installed with YateBTS

Regarding the noise you say, please be more precise, what do you mean? Some noise could be added from your nb AC/DC power supply

[andrew77]

solved!
I forget to create a symbolic link in /var/www to nib web
now path in ybts.conf works

[Vitia]

Hi,

I mean about this test:

http://wiki.yatebts.com/index.php/Calibrate_the_radio_board

1)telnet 127.0.0.1 5038

2) then run the folloing commands

mbts noise

----
and
----

mbts rxgain

(see my last post... the screenshots show, the card is OK)

Which type / gain of antenna are you using?

Vitia

[andrew77]

Hi
just yesterday I make my yatebts working
I tested my bladerf with a RX antenna matched in the downlink band EGSM900 (Whip halfwave dipole, 2dBi) , so with some High VSWR in the UPlink  (with an antenna for tx not matched for this band but far away from the rx antenna in order not to make interference to the rx path): here are the results:

mbts rxgain
current RX gain is 0 dB
mbts noise
noise RSSI is -26 dB wrt full scale
MS RSSI target is -60 dB wrt full scale
mbts rxgain 30
current RX gain is 0 dB
new RX gain is 30 dB
mbts noise   
noise RSSI is --1 dB wrt full scale
MS RSSI target is -60 dB wrt full scale
Instead, with an antenna for RX fully unmatched for this band I got:
 mbts noise
noise RSSI is -9 dB wrt full scale
MS RSSI target is -60 dB wrt full scale
mbts rxgain
current RX gain is 0 dB
mbts rxgain 47
current RX gain is 0 dB
new RX gain is 30 dB
mbts rxgain   
current RX gain is 30 dB
mbts noise
noise RSSI is --3 dB wrt full scale
MS RSSI target is -60 dB wrt full scale
So, It seems that the measurement makes sense. To get the lowest noise one should use a filter taylored for the band in use (the narrow the bandwith, the lowest will be the noise), in addition RX an TX antenna must be far at least  2D^2/lambda (far-field)  where: D=maximum lenght of the antenna and lambda is the lenghtwave of the signal (Ex for 900 MHz lambda =30cm). Next days I will take  better antennas from my work/lab place  to  investigate futher, I will tell you
bye

[Vitia]

Hello,

I see... I have seen this thread @ Nuand: https://nuand.com/forums/viewtopic.php?f=4&t=3578&hilit=http%3A%2F%2Fwww.w1ghz.org

Could you suggest  a low cost (but medium quality -:) ) Amplifier + duplexer for the 900 Mhz band? (max 1 or 2w)

I'm looking for a "safe" solution for the BladeRF board.

Thanks,

Vitia

[andrew77]

A duplexer could be the solution of filtering and, further, amplifing
but 1 or 2W I think is too much to start experimenting

if you are going to buy the http://www.w1ghz.org/small_proj/GVA-84_amplifier.pdf I would suggest to use a traditional "through hole" resistor of around 3 or 4 ohm to feed the monolitic GVA (in the  +5V line), even if this doesn't need.
 How much noise do you have? when:
mbts noise

when I increase the rx gain the noise RSSI increase!

[andrew77]

Look at this I tested it (Made in China) but works well, shipped already tuned.(just to have an idea)
http://shop.sysmocom.de/t/gsm-related/duplexers

http://shop.sysmocom.de/products/dx900-kt30

otherwise:
http://www.mfcsales.com/14540.aspx

[Vitia]

with 2Dbi Antenna SMA male (no filter) oriented in 90 deg:
mbts noise
noise RSSI is -13 dB wrt full scale
MS RSSI target is -50 dB wrt full scale

mbts rxgain
current RX gain is 0 dB

I need a filter before I continue with my test...

[andrew77]

Also good antenna and well matched in the downlink band decreases the noise,
I tested another antenna for RX and increased the noise by 10dB!

Have you ever thought to calibrate DC the board before testing in Yate?
https://github.com/Nuand/bladeRF/wiki/DC-offset-and-IQ-Imbalance-Correction

I never did but I want to try

[Vitia]

Hi,

No, I didn't try that DC calibration. I asked if I could get the filters YateBTS delivered with their kit... I think the board is well calibrated but the filter is really mandatory as it is written in the kit manual... http://wiki.yatebts.com/index.php/File:YateBTSLabKit_QuickGuide.pdf

Do you know this kind of filter? How could I get one such filter? The cavity filter is too big...

Vic

[andrew77]

yes I know,
the kit is provided with the filter,
I looked for you and I have found this:
http://194.75.38.69/pdfs/VBFZ-925+.pdf

Consider that having a narrower bandwidth there are no means but getting a cavity filter,
 otherwise, if you find someone that can make a small pcb you can use this:
http://www.triquint.com/products/p/856671 this is the best solution with low cost!

I did the test with a dummy load connected in the RX and another in the TX port and the result is this:
mbts noise:  noise rssi: -46
so I think this is the practical lowest  noise floor.
If you consider to really buy this filter (it is for Europe) (http://www.triquint.com/products/p/85667)  (there is also the US version) you have to make  a small pcb  and buy two small Female pcb SMA connectors

[Vitia]

Hello,

I'm actually in Quebec (Canada -:) )... so the filter from Miniciruit would be best for me at this time.

If you get that value with dummy load, then there are something wrong in either in the mbts module or in the BladeRF setting. It should be in the -100 area...

Super, I will build the PCB too. I do my pcb here: https://oshpark.com/. It takes time at a very low cost! Gerber or Eagle files work fine.

I see, that you are from the RF field... What I really want to do is to build a Spectrum Analyzer with Tracking generation for Antenna matching work (WiFi 2.4 Ghz band). Do you have a PCB for tracking generator???

Thank you very much!

[andrew77]

well!
I'm from Italy
I  use eurocircuits (provides also Rogers pcb) as pcb maker and Altium Designer for gerbers at my job.
The use of BladeRF as Tracking generator  is one of my desire, but first I want to use the board as fake BTS (imsi catcher)

I used the filter from triquint (or you can find the same product from Epcos) for a GSM/CDMA detector and works very well, but cannot be used for transmitting chain because of its "un-flatness".
The "top" receiver could be done with two of these filters with a very low noise MIMIC amplifier (such as psa4-5043+) between them, but I'm afraid to amplify too much, I don't want to "burn" the front-end of the board; It would need a Variable gain amplifying, etc...
Regarding the mbts noise, I saw in the forums that many folks have these noise values, I will investigate
Eagle pcb can also be used in Linux!
If you need some advice about the pcb don't hesitate to ask
Enjoy!