101+ things to do with a LimeSDR


Very cool indeed! I now desperately want an A-Netz mobile (which I never knew about before) :smiley:


For osmocom-analog: I try to reverse engineer the forgotten “Radiocom 2000”, a French network. Any information, docs or phone donations would help. There are absolutely no standards available on the Internet.


WiFi RADAR to see people moving around in front of their hot spots.

Lightning mapping during a thunderstorm.


30 A Nonlinear Junction Detector (Bug Detector)

I saw this https://hackaday.com/2017/09/20/spy-tech-nonlinear-junction-detectors/ and thought “I wonder if a LimeSDR could do that”. So a quick speed read through one patent from 1971 ( US3631484, there are lots of patents in this area) and it looks like it could possibly work maybe ? It would not be perfect, because you want the TX frequency to be as high as possible so you are limited to one third of 3800MHz (ideally the frequency used should be in a unlicensed ISM/Short Range Devices band). The reason for higher frequency is the larger signal strength returned by smaller devices. Sadly because the two RX channels in a limeSDR can not tune further apart than 61.44MHz this means that only one RX channel can be used, and that it must hop between the 2nd and 3rd harmonic frequency for analysis which is not optimal, but it should still work. It might be slower, but the end result should be the same. Ultimately it just means that you can’t use a Triplexer (LPF,BPF,HPF) to perform the 2nd and 3rd harmonic analysis in real time.

Basic analysis is as follows:

  • If there are no 2nd or 3rd order harmonics present, there is probably little of interest in front of the antenna.
  • If the 2nd harmonic is smaller than the 3rd harmonic, it is probably some metal in front of the antenna.
  • If the 2nd harmonic is larger than the 3rd harmonic, it is probably a nonlinear junction in front of the antenna.

The actual frequency used should not be critical in detecting a NLJ, but higher is going to be a tiny bit more sensitive for extremely small devices.

Frequency to use, and some random numbers for an appropriate diplexer:

Region 1
ISM band
carrier frequency 433.92 MHz (2nd harmonic 867.84 MHz ; 3rd harmonic 1301.76 MHz)
Diplexer: TX 433.92 MHz LPF Cutoff Frequency & RX 867.8 MHz HPF Cutoff Frequency
SRD860 (limited to 25–100 mW ERP )
carrier frequency 866.6 (2nd harmonic 1733.2 MHz ; 3rd harmonic 2599.8 MHz)
Diplexer: TX 866.6 MHz LPF Cutoff Frequency & RX 1733.2 MHz HPF Cutoff Frequency

Region 2
ISM band
carrier frequency 915 MHz (2nd harmonic 1830 MHz ; 3rd harmonic 2745 MHz)
Diplexer: TX 915 MHz LPF Cutoff Frequency & RX 1830 MHz HPF Cutoff Frequency

It is not too hard to design your own Butterworth Lowpass Filters and Butterworth Highpass Filter for a custom made Diplexer.


Just a few numbers:
If your TX power is let say 10mW (+10dBm) then your TX LP harmonic filter should attenuate the 2nd and 3rd harmonic for cca 130dB or you will see this signal in your RX no matter the dipole/diode.

If your TX power is 100mW (+20dBm) then your TX LP harmonic filter should attenate the 2nd and 3rd harmonic for 140dB.

130dB and 140dB is not a trivial task. There is a problem of leakage of the signal over the PCB. Isolation of 140db is very difficult. Even my R&S signal generator start to leak over -131dBm even there is possibility to go down to -144dBm.


OK, so even with a 19th order Butterworth it is not going to achieve 130dB of attenuation probably closer to 100dB

But maybe 19th order Chebyshev Lowpass and Chebyshev Highpass Filter ? I do see your point though, at those kind of signal levels you will have major issues with PCB parasitics and leakage.


I built a harmonic detector some years back that got around the isolation issue using range gating - a harmonic radar.

Two advantages to a range resolved system: first, you can ignore the close-in range bins (where your direct TX/RX term will be) and you get distance to the target.

I used a direct sequence spread spectrum signal at 915 MHz, and looked for the product return at 1830. Going out with one watt I could detect some targets as far away as 100 meters. (Admittedly - that particular target was a diode connected to an antenna, for test purposes).

I was getting circa 60 dB of process gain in the despreading - the receiver MDS was right around -160 dBm in a 10 Hz bandwidth.

I used K&L 9 element cavity filters on TX and RX, and was getting maybe 100 dB isolation. That’s about as far down as it’s possible to measure reliably - the coax you’re using to hook to your instruments only isolates to about that level.


If the DUT (device under test) was 1 meter away from the antenna, I could use the FSPL (Free-space path loss) equation, which at 433MHz would give 25dB signal lost getting from the antenna to the DUT and 25dB for the signals back again. I’ve no idea how inefficient the harmonic generation would be inside the DUT, but I don’t imagine it would be another 25 dB. But in reality the DUT would be 20mm away from the antenna, so it would not be the FSPL equation, it would be a complex near field equation, with direct coupling effects so it would be much much less than 50dB signal lost in total getting there and back. I’m using the FSPL equation because I just want an approximate maximum, (it fails when the distance is less than a few wavelengths). I’m not fully convinced that 130dB of attenuation is strictly necessary, all that is really required is that the returned signal adds enough to the existing harmonics that it’s presence is detectable and measurable. It may require some baseline measurements like pointing the antenna at the sky, hold the antenna to a flat surface, and finally hold the antenna to a flat surface with a diode behind it. I still think that a LimeSDR could be used.


A good target (it was a Nexus 7 tablet, FWIW) had a circa 50 dB conversion loss, incident to 2nd harmonic.

Not so good targets were 80 or 90 dB down. Bad targets were indetectable.


Were any of those tests carried out at a distances less than a wavelength / (2 x pi): 17 millimetres ? (fundamental 915 MHz [0.159 x wavelength = 52.09 mm] ; 2nd harmonic 1830 MHz [0.159 x wavelength = 26.048 mm] ; 3rd harmonic 2745 MHz [0.159 x wavelength = 17.36 mm])

Operating in the “Non-radiative” reactive part of the “near field” region in which there are strong inductive and capacitive effects from the currents and charges in the antenna would produce a higher power transfer.


I tried to operate in the near field, but ran into serious multipath problems so I didn’t pursue it.

My theory re: why the Nexus 7 was such a good target: it was close to one wavelength long in one dimension, and about half that in the other (about 5" x 8"). In the near field coupling is easier, as you indicate - but for my application I wanted large standoff distances.


I do remember some 10 years back (or even more) when I was atending the training in Kongsberg we tested a new technology called Radius https://www.km.kongsberg.com/ks/web/nokbg0240.nsf/AllWeb/57E47945903AB766C12570F3003061A8?OpenDocument

Found some info, where both devices are active (interogator and reflector). This is not what you are looking for…


Convert your front door lock to respond to your keyless entry FOB from your car.

Keyless entry on new cars is very convenient why do we still have keys for our front doors?


This is my goal also. And I want it all to fit in a couple of big ammo cans. I will be using a TI BeagleBoard X15 as my control PC running Linux. I have a long way to go.


My mini is on its way to becoming a small solar powered 7w (25wERP with the collinears) UHF ham repeater.
Once that is running sweet I’ll hack in DMR and every other protocol I can think of.

It’s got the lowest power requirements of any repeater I’ve ever run.


A LimeSDR Mini-based solar powered DMR repeater would be very cool!


31. Signal generator

Although LimeSDR doesn’t seem to be a particularly accurate signal generator, especially on frequencies < 90 MHz, it’s quite possible. See the thread Using LimeSDR as a signal generator


Would like to resurrect harmonic radar using limeSDR.


32. Ground-penetrating radar (GPR)

You would need additional hardware for it to be useful.

I was reading this https://hackaday.com/2019/02/07/ground-penetrating-radar-for-the-masses/ and on page 18 of the linked pdf file I was thinking that at least 20%, possibly even a lot more, of the hardware used in the block diagram could easily be replaced by a single LimeSDR and still be able to generate the exact same frequency sweep from fmin=1.3GHz to fmax=2.6GHz. In a different way but with the same overall end results. In fact that limitation is imposed by the VCO that was used (1300 to 2700 MHz) in combination with a 50 ohm LNA on the receiver (40 to 2600 MHz) input, but if you were using a LimeSDR the sweep range could be increased.