LimeSDR XTRX mPCIe SDR readies for launch!

1 Like

The old Fairwave version had a bandwidth of “120 MSPS SISO / 90 MSPS MIMO” - will the LimeSDR XTRX have such specs?
The Fairwave stack was pitifully full of bugs. My XTRX had never output something else than white noise. What are the progress of LimeMicro for this?

5 posts were split to a new topic: LimeSDR XTRX PCIe Adapter Board

Need to confirm, but I would think so.

The FPGA will have a different gateware, which provides Lime Suite compatibility. Hence it will appear just like any other LimeSDR and be usable from LMS API and gr-limesdr etc.

1 Like

Is the PCB revision the same as the Fairwaves one? Are there any plans to publish the FPGA code?

No, there has been a lot of redesign to make improvements and incorporate replacements for parts which cannot be sourced (nothing that should change performance).

You can see for yourself:

1 Like

I was hoping to use my original XTRX with the new Gateware to use gr-limesdr and Limesuite.
If the FPGA connections with LMS7002M and the PCIe connector are the same it should be possible.

I think the optimisations are mostly related to layout, smaller passives and replacing some EOL parts. Suspect you could use the new gateware, but @Zack could confirm.

It would bei very nice, if even XTRX rev 4 could be used with Limesuite and gr-limesdr, so please check if this could be implemented too, thx

1 Like

Hi @booth,

Our goal is to support original XTRX i.e. we are implementing LimeSuite compatible GW for XTRX rev. 5 boards.

this is very good news for me. I have one XTRX and haven’t been able to get it working yet, the libraries are outdated and buggy.
Litex framework is the best choice!

When I look at the Limesdr comparison table, the sample rate of
LimeSDR 61.44 MSPS
LimeSDR-Mini 30.72 MSPS
LimeSDR-XTRX 30.72

For the PCIe version, I would expect to reach 61.44 MSPS. Where is the bottleneck?


@andrewback @Zack

We also have a XTRX that seems to never have worked and we’d be interested in seeing if the issues come from the software side.

Do you have things that beta testers could start playing with? Or do you have some ETA?

Also, as others have already pointed out, several pages mention a sampling rate of 30.72MSPS, while the original XTRX sported "120 MSPS SISO / 90 MSPS MIMO”. Do you confirm similar sampling rates are to be expected?

1 Like

Yes, sampling rates definitelly will be the same as for original XTRX i.e. "120 MSPS SISO / 90 MSPS MIMO”.

I will let you know when gateware and LimeSuite software will be ready for this. I expect at some point next week.


I’m really excited about this project, because I’ve been wanting an affordable SDR that can integrate on-board like this. I’d also be very interested in seeing an updated version in an M.2 form factor.

1 Like


You mentioned XTRX rev.5 backward compatibility. What about rev.4? They use Xilinx Artix 7 35T. There are a ton of devices out there, including mine. We tried to fix the library issues 2 years ago (if you remember), but it turns out that some of the issues were in the gateware. Would be really nice to support the Xilinx Artix 7 35T based versions as well.

Lime have only recently taken over manufacturing and support for XTRX moving forwards, so I’m guessing you were working with Fairwaves 2 years ago?

I have no affiliation with Fairwaves, I am just a user who owns a rev.4 board. I ported back the XTRX support for upstream gr-osmosdr as the original authors never did that, and eventually that broke compatibility. Besides that we tried to run srs-lte on the XTRX which revealed a massive spectrum flatness issue (I posted many measurements about the triangle shaped LTE FDD carrier with vector analysis if you remember), but that issue was never fixed, although we tried. I suspect there is an uplink issue as well, but due to the DL flatness issue we never get close to diagnose that, although with Osmo-GSM on 2G - where flatness is much less of an issue - I was never able to make it work, and my LimeSDR Mini works fine in both cases of 2G and LTE with the same exact RF IC. So these are likely SW/gateware issues.

Point is, that the older and newer FPGAs on the rev5 and rev4 boards seems to be pretty close, so hopefully the gateware @Zack was talking about would be compatible. I understand it is not a priority, but it would help a lot of XTRX customers who were left without proper device support, and as you can see in this topic, quite a few of them are disappointed.

What I can offer is testing on rev.4 once there is something available to test :slight_smile:

Right, so @Zack and I work for Lime, not Fairwaves, and as you note in your post from last year, Fairwaves had developed their own driver, which they supported. We are in the process of transitioning XTRX over to become a MyriadRF project — it wasn’t when you previously posted and isn’t just yet — and this will mean our own gateware and Lime Suite support.

I’d hope so too, but I’m not familiar with the differences and it may be as simple as that we do not own any rev4 boards to test with, but of course there could be other considerations.

Thanks, I would be interested to hear the results, but caveat emptor of course.

Thanks for the info! Much appreciate it.

I am fully aware that XTRX originally was Firwarve’s own gateware and driver, that is why I am happy that some like you and Zack takes over, meaning we might have a fighting chance to revive these devices.

If there is something to test, I am happy to do so, I can also do signal and vector analysis to verify the output of my rev.4 board. As much as it seems, the two FPGAs are from the same line, the rev.5 has a more powerful version of it. Hopefully they are compatible or it is not too difficult to make it so. @Zack maybe able to comment on that.

Rev.3/4 boards have: Xilinx Artix 7 35T
Rev.5 boards have: Xilinx Artix 7 50T
Just for reference:

1 Like

Thanks for the info @dchard . I just noticed I also have a rev 4 board, not rev 5.

@Zack Any idea if the smaller FPGA of the rev 4 would be enough with the new gateware you’re working on?