-
Notifications
You must be signed in to change notification settings - Fork 136
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
test with limesdr and other sdrs supported by soapysdr #86
Comments
How do I sign up to test with a HackRF? |
Not necessarily cheaper but the usrp b205mini-I works with uhd. Confirmed, at least in my case, that it works. Is there a way to pass a soapy argument to srsue via the crocodile py file? |
I have a lime SDR but would need some guidance to test |
I have the three devices. I will give it a try. Where should I report the results ? |
me too. I have lime SDR and would like to test but would need some guidance |
You will have to start by enabling the SOAPY SDR drivers in https://github.com/EFForg/srsLTE/blob/master/CMakeLists.txt and then get your device working with SOAPY SDR. Compile SRSLTE by hand then test it with one of the test scripts. If it all works please let me know here! |
Made this for someone else trying to use soapy and srslte. I don’t really have an issue using soapy so long as there’s only one device plugged in and/or being recognized by soapy as I don’t know yet how to pick individual devices when using soapy with srslte. It just picks the first device in the list so make sure you have only one thing being recognized by SoapySDRUtil, otherwise if you know the arguments to select specific devices please share. Also, while soapy works when using some things with srslte directly, not sure how to pass this along when starting it with the crocodile file. |
So I have SoapySDR running with the HackRF. It looks like it is found without issue.
I made sure that the SoapySDR was enabled in the CMakeLists.txt file for srsLTE.
When running the crcodilehunter.py I get the following output. Is there something that needs to be passed to tell the script to use SoapySDR?
|
I showed in the video above what needs to be passed when running srslte normally to use soapy. It doesn’t appear this is possible with the crocodile script. |
ah yea the problem here is the bootstrap script I have looks for either a bladeRF or Ettus device and quits if it doesn't find one. If you know bash this should be pretty easy to patch, if not I can try to get a patch today . If you run |
closing #91 as a duplicate of this issue. |
Yeah, I noticed that. I commented the tests out and found an issue with my gps. I'm working on that at the moment. I'll let you know if I get to the point where I can build and test the hackrf with the cell_measurement. Hopefully soon. |
So I was able to get it to run by commenting out the bladeRF and Ettus device checks. The web interface starts but still no joy on the SoapySDR for HackRF.
|
So even if you build in soapy support (I just did with no errors), if you have a look at the examples they only allow or at least say they do “ -a RF args” and not device_name which is what I think is needed. If you have a look at the UE.conf you’ll see that device_name is different than device_args. I’m trying to run any of the examples with soapy but it reverts to uhd. I’ll keep trying but I think something more is needed. |
I don't think that the |
Go to the src/srslte directory. See where it’s failing. Are you building on 20.04? If so, it doesn’t build. 18.04 works. |
I am building on 20.04. So I need to do this on 18.04? I'll get a VM put together and see if I can try that out this afternoon. |
I put together a video of how I built it on a project I maintain, which is based on 18.04. You could maybe use the video as a reference should you encounter the same problems as me. Later today or tomorrow I was going to just upload a dvd that has everything running by default.
But as to 20.04. I could not get the required version of srslte to compile, however, I could get the newest srslte to compile. So I tried various changes mentioned in other issue tickets but no luck for me.
https://youtu.be/4WyJ8oVl4bo
…Sent from my Android
On Aug 17, 2020, at 4:40 PM, phantomdennis ***@***.***> wrote:
I am building on 20.04. So I need to do this on 18.04? I'll get a VM put together and see if I can try that out this afternoon.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or unsubscribe.
|
I used the DragonOS, thank you @alphafox02 and I was able to get the system compiled and running a bit more than before. See output below. This is with the HackRF and SoapySDR. Any thoughts @cooperq?
|
yea the process is crashing but we don't know why, if you run cell_measurement on its own what is the output? |
This is the output:
Then it just sits there. |
So couple things, that exception trace back thing - it’s not actually crashing. It does it on my end and appears to be doing nothing but in fact it’s actually working and will find cells and move on. Each time it goes to look for a cell it’ll pop back up that exception thread 3 thing or more as you can sorta see in the background of the video I posted. As for running cell measurement on its own you need to run it something like this ./cell_measurement -z 900 Or whatever earfcn list you have.. but as I mentioned there’s no way to specify soapy and so I believe it sticks to uhd only which hangs the hackrf. You can change devices args but not the device name. So I think you’ll find that crocodile hunter will work for you if you had a bladerf or usrp. I’m sitting here now using usrp and seeing that same “exception in thread” rubbish you show above. No impact on it running though. |
I also just went ahead and added crocodile hunter to the ISO as that was quicker than making another one. I’ll upload it tonight and hopefully that’ll help some people... but I to would love this soapy + LimeSDR. or whatever. I think the examples need more added though. Hope I’m wrong. |
I just edited the cmake file and disabled uhd and bladerf and only enabled soapy. I’ll see what happens, maybe that’ll let us know if those examples are made to run anything but uhd. |
I'm fairly certain it is using SoapySDR. I adjusted the CMakeList.txt to use only SoapySDR and turned off the UHD and USRP. I also ran the hackrf_info and the device is listed as busy while the cell_measure is running. I'm just not getting any output. What would I get if it found something with cell_measurement? |
I did a cell search. An suggestion on the band? I was thinking 700mhz? I dont know if I need to put this in hz or khz
|
Okay this is looking good at least in my test with an ettus. I deleted the build directory for srslte. I then edited the CMakeLists.txt and put OFF next to uhd and bladerf and on next to soapy. mkdir build Now when I run the cell measurement example it tells me it’s using Soapy and not uhd!! Now the only other issue may be that I have soapy audio installed and it’s listed as the device 0 and my ettus is device 2. So with you running DragonOS do this SoapySDRUtil —find If it’s listing anything other than your hackrf then remove it. Sudo apt-get remove soapysdr0.7-module-audio. |
Nice! We both did the same thing. |
Bingo I can search with HackRF :)
|
I’ll have a lime tomorrow or Monday. Also if you look at this link It shows how we can build work 20.04. Which I’m testing now and hoping to just include in m new DragonOS project. |
LimeSDR mini in hand. It looks promising as it starts and runs with soapy but I’m getting some error about insufficient rx gain etc.. I think I need to get the device args right and manually place them in the crocodile file after I figure out what’s right running the cell measurement tool on its own. I’ll video it all tonight and see if anyone has advice, but again it looks promising at least with my quick check on 20.04 with latest LimeSuite from the PPa. There’s still checking it in 18.04 and with different s versions of soapy etc etc.. hoping it’s just the device args though that’ll make it work regardless of other variables. |
15:00 minutes in I start to mess with the LimeSDR and crocodile hunter. I think it’s a matter of correct device args being passed and if so maybe the crocodile hunter will need to have a way to pass them across to cell measurement, unless you just hard code them in the script. Open to any suggestions anyone may have on what I could do differently to get it working. |
I noticed in your video your clock reference for the LimeSDR is at 40 MHz. In the srsLTE documentation it mentions that devices need to run a clock of 30.72 MHz. You may need to set the clock reference to 30.72 MHz. In the srsLTE issue 437, the user was able to get the clock reference to 30.72 MHz.
|
I’ve found a way to change the clock reference in the lime suite GUI but I think there’s a difference between the normal LimeSDR and the mini. https://discourse.myriadrf.org/t/limesdr-on-srslte-unable-to-find-network-on-lte-device/5182/2 |
Unfortunately any changes applied to the GUI are relevant while the board is connected to it. Any configs in the GUI will go away as soon as you disconnect the board from it. I believe you can still save and restore the configs through the GUI but nothing gets stored to the chip. Edit: |
Not sure if this would help any, but I’ve tried this program with the LimeSDR mini and it seems to work, https://github.com/falkenber9/falcon “ Note: FALCON requires a patched version of srsLTE 18.12 that is automatically downloaded and included as subproject during the build process.” |
I got crocodilehunter working on LimeSDR mini (ubuntu 18.04). I found 3 issues:
Strangely, If I try to run cell_measurement seperately, it will run for a while, locate some CELLs and then crash.
Update: |
I ended up trying the mini again and found that it located one cell when I let it run using uhd soapy vs forcing it to just the straight normal soapy. I would always get an error mx code 4 and something about the tx Calibration which is mentioned a lot on forums. Considering you found cells and I found one I have some hope it’ll be usable with some more work. |
Yes, I also get the
but only once. I fixed the first errors as you said. |
The memory leak error was fixed in the commit referenced in #84. If you pull the latest release (and update the srsLTE sub repo) you will get the fix. |
UPDATE: IT WORKS see next comment I just ran cell_measurement on my machine with a lime SDR mini and also get the error above:
|
Update: IT WORKS
and here is the output from crocoidle hunter :
And it goes on like that finding several other cells. The only change I made was to remove the SDR checking stuff from the bash script. 0ab5ac5 In conclusion try the following steps:
If those don't let you run crocodile hunter with the limeSDR mini please comment here. |
Great work @cooperq ! I had mine running for 5 minutes (indoors) and got I had one last question about the towers I get. I seem to be always getting a "eNodeb id" with value 0 and a PLMN value of 0-0. I get other ledgit towers, but that entry is always there too. |
yea I've noticed that in the past. I've been told that a 0-0-0 cell can be broadcast when an eNodeB is rebooting sometimes. You should go track it down using the GPS and direction finding! |
This is great news. Thanks for making so many improvements. I’ll test it out ASAP. I’ve been using usb3 and good antennas so hopefully the improvements in the code will yield good results. A mini is such a great priced radio to use for this project! |
On another subject, MyriadRF confirmed here that they're working on a native version of the LimeSDR library for srsLTE which will also (hopefully) make things much smoother. |
Speaking of uhd and soapy.. Are you guys now testing and finding good results with just soapy or uhd soapy when running the crocodile hunter? edit: I see the outputs above which look as though you’re using just straight soapy. I’ll try both. |
I was using just soapy with the lime rf. I'll try UHD Soapy as well with the UHD device I have though. |
I just mentioned uhd soapy because if I build this and have the uhd driver and soapy installed uhd Soapy takes precedence with the lime sdr until I go disable uhd in the cmake file and build srslte with just soapy support. |
does this cause a bug? Can you show me what the output is? I'm a little slow today but I don't totally understand the problem. |
Built the latest and it works great so far with the lime-sdr. I didn’t change a thing at first so when I started it up it was using the “uhd soapy” with the LimeSDR. It seemed to run but something didn’t seem right, didn’t run it long enough to really evaluate it well. So I stopped it, went in to srsLTE and changed the CMakefile list file to disable uhd and enable soapy. Rebuilt and reran and within a few minutes the LimeSDR found two cells/towers. It looked like it was running way faster with normal soapy when compared to uhd soapy. Thanks for making this possible! |
This is with the CMakelist file like this, which I had to manually do and it seems to run faster with the limesdr. option(ENABLE_GUI "Enable GUI (using srsGUI)" OFF)
And with it like this, which is how it is by default (I'm running it again this way now and will get you feedback asap) option(ENABLE_GUI "Enable GUI (using srsGUI)" OFF)
|
I know this is now closed. Just wanted to let you know that with the latest "master" checkout of LimeSuite, I don't get the " [ERROR] Tx Calibration: MCU error 4 (SXT tune failed)" any more |
Oh interesting. I’m using the packaged version in 20.04. Should I maybe look at replacing that and build with source? |
These latest changes in the LimeSuite repo were probably triggered by the following MyriadRF forums discussion: https://discourse.myriadrf.org/t/how-to-fix-the-error-tx-calibration-mcu-error-4-sxt-tune-failed/ I would certainly recommend compiling from source. You just need to make sure that you also compile the latest 0.7.x version of Soapy SDR as the Master SoapySDR version defaults to 0.8 which is not supported by LimeSuite yet. tip: I used checkinstall to install the compiled versions instead of doing a make install as it generates deb files and it's cleaner. |
Thanks for the tip! I do have soapysdr 0.7 in place and I like the idea of the deb packages, I’ll have to give it a try ASAP. |
Thanks for coninuing to test! |
Test to see if crocodile hunter can run on a cheaper SDR:
The text was updated successfully, but these errors were encountered: