Re: [Openbts-discuss] OpenBTS bladeRF support

[Robert G. <rob...@nu...>]

Ralph,
  Would you mind checking out the latest changes in
https://github.com/Nuand/OpenBTS/commit/e9a60f16bf101a01e13d803ba61b627893869b60?
If you are using my dev checkout you can run ./pull.sh from the main
directory.
  I'm hoping you'll see significantly better performance :)

Alexander,
  Sorry about not responding to this sooner, but what I do is I dump the
RACH bursts from the transceiver and import them into Matlab and perform
FFTs+EVMs. It's kind of hacky but it works.


Rob



On Fri, Apr 11, 2014 at 2:57 AM, Alexander Chemeris <
ale...@gm...> wrote:

> Same here with E4406A's. How do you test Rx quality?
>
> Please excuse typos. Written with a touchscreen keyboard.
>
> --
> Regards,
> Alexander Chemeris
> CEO/Founder Fairwaves LLC
> https://fairwaves.co
> 11 апр. 2014 г. 12:15 пользователь "Robert Ghilduta" <
> rob...@nu...> написал:
>
> Ralph, yea sorry about mixing up the email thread there! The gain values
>> for RX are hardcoded in the openbts/TransceiverRAD1/bladeRFDevice.ccp file
>> at line 240 & 241, here's a link
>> https://github.com/Nuand/OpenBTS/blob/master/TransceiverRAD1/bladeRFDevice.cpp#L240.
>>
>> Alexander,
>>   We have a few radios hooked directly to a few E4406As and have been
>> calibrating our output signal using those. Thanks for the info, I will test
>> to see what happens if setTxGain() toggles txvga1 and let you know!
>>
>> Rob
>>
>>
>>
>> On Fri, Apr 11, 2014 at 12:18 AM, Alexander Chemeris <
>> ale...@gm...> wrote:
>>
>>> Robert,
>>>
>>> For UmTRX we found that the best TxVGA1 value is around -21dB. TxVGA2
>>> doesn't have much influence, unless you drive it too high.
>>>
>>> What setup do you use for testing Tx and Rx quality of the transceiver?
>>>
>>> Please excuse typos. Written with a touchscreen keyboard.
>>>
>>> --
>>> Regards,
>>> Alexander Chemeris
>>> CEO/Founder Fairwaves LLC
>>> https://fairwaves.co
>>> 10 апр. 2014 г. 22:07 пользователь "Robert Ghilduta" <
>>> rob...@nu...> написал:
>>>
>>>  Thanks for the update Ralph! Currently our RX and TX gains are
>>>> hardcoded. There's a bit of work needed to figure out how to tune the
>>>> VGA/PA/LNAs in the LMS as to maximize SNR but we'll get there!
>>>>
>>>>
>>>> On Thu, Apr 10, 2014 at 11:02 AM, Ralph A. Schmid, dk5ras <
>>>> ra...@sc...> wrote:
>>>>
>>>>> Update: Voice works, even with latest version - but not on first try,
>>>>> usually at least the first call after boot is a muted call...
>>>>>
>>>>>
>>>>>
>>>>> And chans reports a for 1 dB different downlink level as shown in the
>>>>> network monitor. I have no idea wether the netmon is wrong, the reporting
>>>>> is wrong or OpenBTS is wrong J
>>>>>
>>>>>
>>>>>
>>>>> Ralph.
>>>>>
>>>>>
>>>>>
>>>>> *From:* Ralph A. Schmid, dk5ras [mailto:ra...@sc...]
>>>>> *Sent:* Thursday, 10 April, 2014 05:26
>>>>> *To:* 'Michael Iedema'
>>>>>
>>>>> *Cc:* ope...@li...
>>>>> *Subject:* Re: [Openbts-discuss] OpenBTS bladeRF support
>>>>>
>>>>>
>>>>>
>>>>> Hmmm, I thought it should not be an issue with stand-alone systems out
>>>>> of the box?! Is it documented somewhere what entry I have to add?
>>>>>
>>>>>
>>>>>
>>>>> Ralph.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> *From:* Michael Iedema [mailto:mic...@ra...<mic...@ra...>]
>>>>>
>>>>> *Sent:* Wednesday, 9 April, 2014 23:13
>>>>> *To:* Ralph A. Schmid, dk5ras
>>>>> *Cc:* Robert Ghilduta; ope...@li...
>>>>> *Subject:* Re: [Openbts-discuss] OpenBTS bladeRF support
>>>>>
>>>>>
>>>>>
>>>>> Could be the new GSM.CallerID.Source setting in OpenBTS.
>>>>>
>>>>>
>>>>> On Apr 9, 2014, at 9:47 PM, "Ralph A. Schmid, dk5ras" <
>>>>> ra...@sc...> wrote:
>>>>>
>>>>> Something in the master repo from range killed voice calls - I need to
>>>>> revert to my backup to restore this. Did not yet sort out what packages
>>>>> fault it was, maybe I will find time tomorrow. Somehow difficult when two
>>>>> different versions are used in one VM J
>>>>>
>>>>>
>>>>>
>>>>> Ralph.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> *From:* Ralph A. Schmid, dk5ras [mailto:ra...@sc...<ra...@sc...>]
>>>>>
>>>>> *Sent:* Wednesday, 9 April, 2014 19:40
>>>>> *To:* 'Robert Ghilduta'; ope...@li...
>>>>> *Subject:* Re: [Openbts-discuss] OpenBTS bladeRF support
>>>>>
>>>>>
>>>>>
>>>>> Congratulations! I am logged on with some phone, first GPRS packets
>>>>> flow... I did  not change anything, but at home it suddenly works JNeeded some time to settle and rethink?!
>>>>>
>>>>>
>>>>>
>>>>> Phones can call each other, they ring, but no voice yet. Have to test
>>>>> against it with SDR1 if it is an asterisk issue or something else.
>>>>>
>>>>>
>>>>>
>>>>> Ralph.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> *From:* Robert Ghilduta [mailto:rob...@nu...<rob...@nu...>]
>>>>>
>>>>> *Sent:* Wednesday, 9 April, 2014 16:44
>>>>> *To:* Ralph A. Schmid, dk5ras
>>>>> *Subject:* Re: [Openbts-discuss] OpenBTS bladeRF support
>>>>>
>>>>>
>>>>>
>>>>> Hoopycat's images are up-to-date and can be used in place of the ones
>>>>> found in ./openbts/TransceiverRAD1/
>>>>>
>>>>> We haven't done aging analysis of the VCTCXO just yet but after 6
>>>>> months the thing is likely to drift a few ppm. We have kalibrate support in
>>>>> the works so then the VCTCXO can be calibrated against GSM basestations.
>>>>>
>>>>>
>>>>>
>>>>> On Wed, Apr 9, 2014 at 7:39 AM, Ralph A. Schmid, dk5ras <
>>>>> ra...@sc...> wrote:
>>>>>
>>>>> The LED flashes, the network is not found by my phone, but maybe just
>>>>> the frequency is too much off. At home I have the test equipment to see if
>>>>> it transmits. Can the TCXO be permanently calibrated? Then I will align it
>>>>> with my Rubidium Atomic Clock, so the frequency should be accurate enough
>>>>> J
>>>>>
>>>>>
>>>>>
>>>>> Ralph.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> *From:* Robert Ghilduta [mailto:rob...@nu...]
>>>>> *Sent:* Wednesday, April 9, 2014 4:14 PM
>>>>>
>>>>>
>>>>> *To:* Ralph A. Schmid, dk5ras
>>>>> *Cc:* ope...@li...
>>>>> *Subject:* Re: [Openbts-discuss] OpenBTS bladeRF support
>>>>>
>>>>>
>>>>>
>>>>> Awesome, looking forward to hearing how it works out for you!
>>>>>
>>>>> I added all of the bladeRF specific build steps to our GitHub fork at
>>>>> https://github.com/nuand/dev .
>>>>>
>>>>>
>>>>>
>>>>> PS. Just make sure to upgrade the flash RBF image to one of the RBFs
>>>>> found in ./openbts/TransceiverRAD1/
>>>>>
>>>>>
>>>>>
>>>>> Rob
>>>>>
>>>>>
>>>>>
>>>>> On Wed, Apr 9, 2014 at 7:06 AM, Ralph A. Schmid, dk5ras <
>>>>> ra...@sc...> wrote:
>>>>>
>>>>> OK, this was the missing bit - I already thought so that there must be
>>>>> some option to set :)
>>>>>
>>>>>
>>>>>
>>>>> FPGA is loaded from the SPI FLASH automatically, this is a feature I
>>>>> really love in the bladeRF.
>>>>>
>>>>>
>>>>>
>>>>> I will try later and let you know!
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> Ralph.
>>>>>
>>>>>
>>>>>
>>>>> *From:* Robert Ghilduta [mailto:rob...@nu...]
>>>>> *Sent:* Wednesday, April 09, 2014 3:24 PM
>>>>>
>>>>>
>>>>> *To:* Ralph A. Schmid, dk5ras
>>>>>
>>>>> *Cc:* ope...@li...
>>>>>
>>>>>
>>>>> *Subject:* Re: [Openbts-discuss] OpenBTS bladeRF support
>>>>>
>>>>>
>>>>>
>>>>> Hi guys,
>>>>>
>>>>>   Sorry about the delay, and thanks for the warm welcome!
>>>>>
>>>>>   Michael, our initial development was done against r6817 but I
>>>>> recently upgraded it to follow the OpenBTS repository found in
>>>>> RangeNetwork's Github. I can certainly create a pull request, I would just
>>>>> like to wait a bit to make sure all of the bugs have been worked out!
>>>>>
>>>>>
>>>>>
>>>>> Ralph,
>>>>>
>>>>>   USB2.0 support still has some issues to overcome, so at the moment
>>>>> USB3.0 is mandatory.
>>>>>
>>>>>   If you grab the latest firmware and FPGA images from
>>>>> http://hoopycat.com/bladerf_builds/ you should be in good shape.
>>>>> Currently the transceiver does not load the FPGA so you have to manually
>>>>> load the FPGA image with the correct RBF from the TranasceiverRAD1/
>>>>> directory or from the previously linked site.
>>>>>
>>>>>   As for the build issue, if you run `export confflags=--with-bladeRF`
>>>>> before doing ./build.sh the build scripts should switch over to building
>>>>> bladeRF devices. Running `grep bladeRF openbts/TransceiverRAD1/transceiver`
>>>>> should hopefully say the "binary file transceiver matches".
>>>>>
>>>>>
>>>>>
>>>>>   I also have the tendency of running compiling and running openbts
>>>>> directory from the build files in the git checkout directory. Once
>>>>> everything is up and running, LED1 and LED3 should be solid greed, while
>>>>> LED2 should start to blink at ~4Hz. Let me know if you see something
>>>>> similar.
>>>>>
>>>>>
>>>>>
>>>>> Rob
>>>>>
>>>>> On Wed, Apr 9, 2014 at 4:32 AM, Ralph A. Schmid, dk5ras <
>>>>> ra...@sc...> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>>
>>>>>
>>>>> Even after removing my OpenBTS V4 standard installation and creating a
>>>>> new dev folder from your repo, following the standard procedure (git clone,
>>>>> clone.sh, pull.sh, build.sh, dpkg), it looks to me that my PC alway builds
>>>>> a normal rangenetworks hardware installation.
>>>>>
>>>>>
>>>>>
>>>>> Is this possible, is there some step I am missing to point the
>>>>> procedure to build a bladeRF installation?
>>>>>
>>>>>
>>>>>
>>>>> Ralph.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> *From:* Robert Ghilduta [mailto:rob...@nu...]
>>>>> *Sent:* Monday, April 07, 2014 4:30 PM
>>>>> *To:* ope...@li...
>>>>> *Subject:* [Openbts-discuss] OpenBTS bladeRF support
>>>>>
>>>>>
>>>>>
>>>>> Hello,
>>>>>
>>>>>    We have been working on adding bladeRF support to OpenBTS for quite
>>>>> a bit now. Currently the transceiver is built upon TransceiverRAD1. The
>>>>> bladeRFDevice relies on the libbladeRF library that provides it with a
>>>>> means of doing asynchronous buffering and timestamping with the help of the
>>>>> FPGA. The system seems to boot up and beacon just fine, and I can even see
>>>>> AGCH bursts and AUTH requests when trying to associate with an MS in
>>>>> OpenRegistration mode however the MS (a Nokia 3310) always displays "No
>>>>> access" after a few seconds of attempting to associate.
>>>>>
>>>>>    My OpenBTS.log ( http://nuand.com/captures/OpenBTS.log ) shows
>>>>> that its able to progress pretty far into the association process. I've
>>>>> also captured a few GSMTAP packet captures of a typical association,
>>>>> http://nuand.com/captures/gsmtap.pcapng .
>>>>>
>>>>>    Is this there anything here that could be the underlying
>>>>> timestamping mechanisms fault? Could it be a bad configuration or faulty MS?
>>>>>
>>>>>
>>>>>
>>>>>    For anyone that is interested in experimenting or debugging with
>>>>> what we have so far, you can take a look at my cloned dev repository at
>>>>> https://github.com/robertghilduta/dev . My last two commits provide a
>>>>> mechanism for overriding the location from which certain repositories are
>>>>> fetched. In this case OpenBTS is fetched from my repository (
>>>>> https://github.com/robertghilduta/openbts ).
>>>>>
>>>>>
>>>>>
>>>>> Thanks,
>>>>>
>>>>>   Rob
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>>
>>>>> Robert Ghilduta
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>>
>>>>> Robert Ghilduta
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>>
>>>>> Robert Ghilduta
>>>>>
>>>>> Nuand LLC, Co-Founder
>>>>>
>>>>> Tel: 1-917-837-1706
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>>
>>>>> Robert Ghilduta
>>>>>
>>>>> Nuand LLC, Co-Founder
>>>>>
>>>>> Tel: 1-917-837-1706
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Put Bad Developers to Shame
>>>>> Dominate Development with Jenkins Continuous Integration
>>>>> Continuously Automate Build, Test & Deployment
>>>>> Start a new project now. Try Jenkins in the cloud.
>>>>> http://p.sf.net/sfu/13600_Cloudbees
>>>>>
>>>>> _______________________________________________
>>>>> Openbts-discuss mailing list
>>>>> Ope...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/openbts-discuss
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Put Bad Developers to Shame
>>>>> Dominate Development with Jenkins Continuous Integration
>>>>> Continuously Automate Build, Test & Deployment
>>>>> Start a new project now. Try Jenkins in the cloud.
>>>>> http://p.sf.net/sfu/13600_Cloudbees
>>>>> _______________________________________________
>>>>> Openbts-discuss mailing list
>>>>> Ope...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/openbts-discuss
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Robert Ghilduta
>>>> Nuand LLC, Co-Founder
>>>> Tel: 1-917-837-1706
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Put Bad Developers to Shame
>>>> Dominate Development with Jenkins Continuous Integration
>>>> Continuously Automate Build, Test & Deployment
>>>> Start a new project now. Try Jenkins in the cloud.
>>>> http://p.sf.net/sfu/13600_Cloudbees
>>>> _______________________________________________
>>>> Openbts-discuss mailing list
>>>> Ope...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/openbts-discuss
>>>>
>>>>
>>
>>
>> --
>> Robert Ghilduta
>> Nuand LLC, Co-Founder
>> Tel: 1-917-837-1706
>>
>


-- 
Robert Ghilduta
Nuand LLC, Co-Founder
Tel: 1-917-837-1706