LimeSDR running on HF receive today

This morning, I quickly assembled one of my KN0CK HF Upconverter modules from one I had on hand (version 4 that I designed/built about 2 years ago), stuck it ahead of the LimeSDR, and the results were REALLY amazing…! See for yourself with this series of videos that I published to Youtube this morning. Until the LimeSDR software apps are fully developed to achieve operation on the HF band, this should give everyone a pretty fair idea how it’ll work - and if not, the fallback to use Up/Downconversion will yield pretty nice results when used with the LimeSDR. Here are the videos I took of various modes of operation and bands during Amateur Radio’s Field Day today, too:

The same videos have been blogged today, too - So the LimeSDR is getting coverage again… :slightly_smiling:

73 de Marty, KN0CK


What receive filters do you have on the LimeSDR board?


I’m just using what’s there already and applied to the LNAL port. I just assumed that the LimeSDR was originally optimized for much higher frequency operation and just looking at the design I would think that it’s going to be pretty optimal in the UHF/SHF range (since it was originally going to be used for Wifi and Cellular radio) but it seems to have pretty impressive performance into the VHF range, too. I’ve had my KN0CK HF Upconverter (120MHz) attached to the LNAL port on my Lime and the 120 MHz upconversion and resulting demod from the Lime from the HF band is IMPRESSIVE…! On 20m it’s every bit as good (God, if not better) than my Flex 5000A in terms of receive quality. No kidding, it’s impressive.

I would think that one of the HW developers at Lime Microsystems would know the input filter characteristics enough to tell us what it’s capable of. Just guessing, I found that it’ll take the 120 MHz upconversion from my HF Upconverter and does a marvelous job with it - the video is just the icing on the cake.

73 de Marty, KN0CK

I haven’t checked the schematics but from what I see on the PCB I don’t believe there are any input filters. The parts on the PCB look more like wide band transformes used for impedance matching and / or to convert between single ended and differential.

I have also run some “fake HF” tests some time ago using an upconverter and and it worked very well, see the screenshots below :slightly_smiling:


That looks INCREDIBLE…!! If the input circuits are capable of HF then your results you show in the upconverted HF band (like I did today) will look just like that. God, I cannot wait for some of this software to mature…!

Keep on it…I’m designing the translator for the Lime right now - a lot of it was a ‘lift and drop’ exercise from my RTL-SDR Upconverter design with just the transmit Downconverter added. it may not be necessary in the end when all the software matures, but I’m hedging a bet with myself for now to try this idea.

73 de Marty, KN0CK

The upconverter results look very promising! Can’t wait for someone to demonstrate the direct HF operation. Thanks, guys!

Thanks very much for the post, Chuck…All I can say is that if you got in with the LimeSDR campaign early (as an Early Bird 1 or 2 Batch), you WILL NOT be disappointed - the LimeSDR is a phenomenal receiver…! I would put it up against my Flex 5000 now and can tell you that the LimeSDR can hear things that the Flex struggles to deliver.

Also, some of the most brilliant minds are working the project - people like Josh Blum, Alex Csete, and Simon Brown among many more at Lime Microsystems. It’s like the dream team for SDR development…So what will be there in November should be incredible.

Stay tuned for yours in November…73,

de Marty, KN0CK

It’s very sensitive, that’s for sure :slight_smile: .

There are no filters on the inputs and, as you say, only wideband baluns.

Also looks as though @joshblum may have now added support for the onboard NCO and native tuning for HF bands:

FWIW there are filters on my boards in accordance with the schematics.


On another thread related to the connectors used on the Lime board, Ricardas mentioned the following (his direct comments):

"…LMS7002M has 3x Rceiver RF inputs and 2x Tx output per channel. By design receiver inputs are tuned as follows:

LNAL - from 0.1 MHz to 2000 MHz. Gives a noise figure <2dB.
LNAH - from 1500 MHz to 3800 MHz. Gives a noise figure <3dB.
LNAW - from 0.1 MHz to 3800 MHz. Gives a noise figure 5-7dB. Usually used for sniffing.
Transmitter output bu design are similar and supported frequency range is form 0.1-3800 MHz.

However, every RF input/output is combined with RF matching network which might be band specific. On the LimeSDR the RF inputs/outputs are tuned for:

LNAL - is tuned for 800 - 900 MHz (3GPP Band V).
LNAH - is tuned for 1900 MHz (3GPP Band I).
LNAW - has broadband matching up to 2000 MHz.
TXout1 - is tuned for 2600 MHz (3GPP Band VII) .
TXout2 - has broadband matching up to 2000 MHz.
The matching network includes broadband Balun + network for fine tune for specific band, hence can be easily modified.


The LMS7002M synthesizer frequency range is from 30 MHz - 3800 MHz. We are using internal NCO to get us to 100 kHz region.


So there’s some REAL nuggets of info for all of us in there…Appears for HF work, LNAL is best used for receive input because that’s how it’s designed for the '7002, but it appears to be optimized for the 3GPP band (800 to 900 MHz) on the board – at least it appears that way in his comments. LNAW appears to have broadband matching up to 2 GHz, but has too much noise figure to be useful for anything else but sniffing.


It’s my understanding from some other posts that the LimeSDR units we are to receive do not have the filtering on the ports so that we can use them as we please instead of being locked into GSM/LTE/WiFi bands on all but one of the ports.

What I am really excited to try out is getting it running with a transceiver program that allows for multiple RX across bands. watching 50M while working 70M or when the HF is worked out working both 40 and 30 or 20 from the same interface will be awesome.

Bandwidth is too low to allow working 2M/70cm while working 20M, but these are cheap enough that I bought two and small enough to putthem in the same housing so that isn’t going to be an issue, in fact it will give me 4 TX/RX to play with. Now I only need to locate the software that can do that…

1 Like